{ "directory": "The Agents Index", "url": "https://theagentsindex.com", "generatedAt": "2026-07-24T15:23:44.874Z", "count": 52, "nextCursor": null, "listings": [ { "slug": "11x", "name": "11x", "tagline": "AI \"digital workers\" for sales, Alice runs autonomous outbound; Julian is an AI voice agent.", "url": "https://theagentsindex.com/11x", "websiteUrl": "https://www.11x.ai", "tier": "Listed", "categories": [ "sales-marketing-agents" ], "tags": [ "enterprise", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "11x is one of the most visible bets in the \"AI SDR\" category: Alice runs the whole outbound motion, research, list-building and multi-channel sequencing, while Julian adds an AI voice agent for calls, both pitched as always-on digital workers that plug into your CRM. Backed by a16z and Benchmark, it’s genuinely capable and the underlying category is real. But the caveats are significant and well-documented: pricing is custom and sales-led with no free plan, autonomous SDR quality varies enough that independent tests report mixed results, and in March 2025 TechCrunch reported the company had displayed logos of firms that said they weren’t customers, alongside former-staff accounts of high early churn, issues that preceded a leadership change. Pick 11x if you’re an enterprise team ready to run a rigorous paid pilot and measure deliverability and booked meetings yourself; skip it if you want transparent pricing, a free trial, or to buy on marketing claims rather than your own data.", "lastVerified": "2026-07-23" }, { "slug": "ada", "name": "Ada", "tagline": "A Toronto-founded, $1.2B-valued AI customer-service agent platform predating the LLM wave — sales-led, no public pricing, a genuine public API.", "url": "https://theagentsindex.com/ada", "websiteUrl": "https://www.ada.cx", "tier": "Listed", "categories": [ "support-agents" ], "tags": [ "enterprise", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "Ada is the incumbent of this category — founded in 2016, well before Decagon, Crescendo and Cresta existed, and still backed by a $1.2B valuation and $190M+ raised. Its strongest genuine differentiator today is developer access: a documented Conversations/Email API and an MCP server give it a self-serve integration path that Decagon, Sierra and Crescendo simply don't offer. The honest catch is the gap between how Ada looks from the admin seat and how it feels from the customer seat — a 4.6/5 G2 rating from the teams who configure it sits well above a 2.0/5 Trustpilot rating from the people actually chatting with it, with real complaints about lost context and reaching a human. Pick Ada if you want an established vendor with real API access and can budget for a sales-led, several-month implementation; skip it if you want transparent pricing, a free trial, or you're wary of that admin-vs-end-user satisfaction gap.", "lastVerified": "2026-07-23" }, { "slug": "aider", "name": "Aider", "tagline": "AI pair programming in your terminal, an open-source CLI agent that edits your code and commits to git.", "url": "https://theagentsindex.com/aider", "websiteUrl": "https://aider.chat", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "self-hosted", "claude", "open-source", "open-models", "free-tier" ], "apiAvailable": false, "pricing": "free", "verdict": "Aider is the definitive terminal AI pair programmer: open-source under Apache-2.0, model-agnostic, and built around a git-native workflow where every AI edit becomes a reviewable, revertible commit. That commit discipline plus its repository map and auto lint/test fixing make it one of the safest ways to let a model touch a real codebase, and being bring-your-own-key keeps it free and flexible — route work to Claude, GPT, DeepSeek or a local model as you like. Its well-known code-editing leaderboards are a genuinely useful, oft-cited signal on model ability. The trade-offs are that it is terminal-only with a steeper on-ramp than a GUI tool, you pay your provider per token, and it is an interactive pair programmer rather than a fire-and-forget agent. Pick Aider if you live in the terminal and want free, auditable, model-agnostic AI edits under git; skip it if you want a graphical editor or a fully autonomous cloud agent.", "lastVerified": "2026-07-23" }, { "slug": "artisan", "name": "Artisan", "tagline": "Ava, the AI BDR, an autonomous sales agent that sources leads, runs multi-channel outbound and books meetings without a human writing each message.", "url": "https://theagentsindex.com/artisan", "websiteUrl": "https://www.artisan.co", "tier": "Listed", "categories": [ "sales-marketing-agents" ], "tags": [ "enterprise", "autonomous" ], "apiAvailable": false, "pricing": "paid", "verdict": "Artisan's Ava is one of the more complete \"AI BDR\" plays on the market: since the May 2026 Ava 2.0 rebuild she runs the full outbound motion autonomously, sourcing, enrichment, multi-channel outreach, reply and objection handling, and meeting booking, with a native dialer bridging the loop to a human for calls. The company is well-capitalised (~$37M+ raised, led by Glade Brook Capital with HubSpot Ventures and Y Combinator) and backs its pitch with specific, published customer numbers rather than vague praise. But two things temper the recommendation: pricing is entirely opaque (every tier is \"Talk to sales,\" and the self-serve entry tier the company announced at launch isn't visible on the live site), and Artisan has racked up a genuine string of controversies in a short window, a viral \"Stop Hiring Humans\" campaign, a three-week LinkedIn ban over a data vendor's ToS violation, and a publicized (since-settled) dispute over unauthorised use of a cartoonist's artwork. Pick Artisan if you're a funded sales team ready to run a proper paid pilot and judge Ava on your own reply and booking data; look elsewhere if you need transparent self-serve pricing or want to avoid a vendor whose brand has drawn this much recent scrutiny.", "lastVerified": "2026-07-23" }, { "slug": "bland-ai", "name": "Bland AI", "tagline": "Enterprise voice AI that runs on its own infrastructure, build calls as conversational pathways, priced all-in per minute.", "url": "https://theagentsindex.com/bland-ai", "websiteUrl": "https://www.bland.ai", "tier": "Listed", "categories": [ "voice-agents" ], "tags": [ "self-hosted", "enterprise", "api", "no-code", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "paid", "verdict": "Bland AI’s pitch is ownership: it runs its own self-hosted model stack instead of stitching together third-party providers, which underpins a genuinely strong compliance story (SOC 2, HIPAA, PCI DSS, EU/APAC residency, on-prem) and an all-in per-minute price with no confusing model pass-throughs. Its Conversational Pathways give the deterministic, auditable control that regulated call flows demand. Be clear-eyed about the limits, though: the marketed 400ms latency and 40+-language support are vendor-reported without independent benchmarking, and support leans on the public Discord community and self-service below the Enterprise tier. Pick Bland if you are an enterprise that values infrastructure control, compliance and predictable pricing above all; skip it if you want a fast, no-code, multilingual agent with hands-on support at a starter price.", "lastVerified": "2026-07-23" }, { "slug": "bolt-new", "name": "Bolt.new", "tagline": "StackBlitz's prompt-to-app builder that runs a real Node.js environment inside the browser tab itself, no remote VM.", "url": "https://theagentsindex.com/bolt-new", "websiteUrl": "https://bolt.new", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "browser-use", "enterprise", "no-code", "autonomous", "free-tier" ], "apiAvailable": false, "pricing": "freemium", "verdict": "Bolt.new’s WebContainers runtime is a genuine, verifiable technical edge — it runs Node.js natively inside the browser sandbox instead of proxying to a remote VM, which is why projects boot and preview instantly with zero local setup. Recent additions (MCP server connectors, Figma import, AI image editing) keep it moving fast, and a real free tier plus published pricing mean you can try it without a sales call. The honest catches: effort-based billing can run up fast on hard debugging loops, there is no public API to drive it from your own code, and its Trustpilot record (1.5/5 from 189 reviews) points to a real, unresolved customer-support problem worth weighing before you pay. Pick Bolt.new if you want the fastest prompt-to-deployed-app path and are comfortable self-serving through any support issues; pick Replit Agent instead if you want a similar zero-setup experience with a more conventional cloud-VM backend, or Cursor/Windsurf if you want an agent inside your own local toolchain rather than a hosted app factory.", "lastVerified": "2026-07-23" }, { "slug": "claude-code", "name": "Claude Code", "tagline": "Anthropic’s terminal-native coding agent with full-repository context.", "url": "https://theagentsindex.com/claude-code", "websiteUrl": "https://www.anthropic.com/claude-code", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "claude", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "Claude Code is the pick for developers who want an agent that lives in the terminal and composes with real engineering workflows, scripts, CI pipelines, headless automation, rather than a bolted-on chat panel. Its reasoning holds up on genuinely multi-step tasks across a full repository, and it asks permission on sensitive actions instead of working blind. There is no graphical editor, so anyone who wants point-and-click review will feel the gap, and usage-based cost on long autonomous runs is worth watching. It is the strongest option for CI-driven or command-line-first teams, and a weaker fit for developers who want a GUI-first experience. One more factor worth knowing: Claude Code itself passed $1B in run-rate revenue within six months of its public launch, and Anthropic, now backed by a $65B Series H at a $965B valuation, is investing directly in its toolchain, having acquired the Bun JavaScript runtime Claude Code ships as an executable in December 2025. The tradeoff of that single-vendor stack is that Claude Code is tied exclusively to Claude models, with no built-in switch to GPT or Gemini the way some editor-based rivals allow.", "lastVerified": "2026-07-24" }, { "slug": "clay", "name": "Clay", "tagline": "GTM data platform with an AI research agent (Claygent), enrich, research and personalise outbound at scale.", "url": "https://theagentsindex.com/clay", "websiteUrl": "https://www.clay.com", "tier": "Listed", "categories": [ "sales-marketing-agents" ], "tags": [ "claude", "api", "gpt", "no-code", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Clay has become the GTM data layer serious outbound teams build on, and for good reason: its waterfall enrichment posts materially higher match rates than any single-source tool, and Claygent adds AI research that scrapes the web for data points no provider packages. Put together with signals, scoring, AI copy, a sequencer and ad-audience sync, it takes you from raw list to CRM-ready outbound in one workspace, with unlimited seats on every plan. The tradeoffs are equally well documented: the dual credit meter is genuinely hard to predict, failed lookups still cost credits and Claygent-heavy runs can balloon at scale, and the tool has a steep enough learning curve that it really wants a dedicated GTM engineer. Pick Clay if you have the technical appetite and a budget above a few hundred dollars a month; skip it if you need predictable flat pricing or have no one to build the workflows.", "lastVerified": "2026-07-23" }, { "slug": "cline", "name": "Cline", "tagline": "The open-source coding agent, an autonomous, bring-your-own-key agent that lives in your editor, terminal or SDK.", "url": "https://theagentsindex.com/cline", "websiteUrl": "https://cline.bot", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "self-hosted", "api", "open-source", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Cline is the reference open-source coding agent: Apache-2.0, bring-your-own-key, and available as a VS Code extension, a CLI and an embeddable SDK, with Zed/Neovim reachable via the open Agent Client Protocol. Its Plan/Act workflow and first-class MCP support make it both careful and unusually extensible, and because you supply the model you get full choice and transparent, usage-only cost — route cheap tasks to a local model and hard ones to a frontier model. Two optional paid layers now sit on top of that free core: a $9.99/month ClinePass for discounted open-weight-model rate limits, and a custom-priced Enterprise tier that adds an SLA, dedicated support and a centrally-managed JetBrains rollout for organizations — the JetBrains plugin itself is free and self-serve, same BYOK terms as VS Code. The flip side of BYO-key is that costs are unbounded on the core tier: a long agentic run on a top model can quietly rack up API spend, and you own all the setup, keys and troubleshooting yourself. Output quality is only as good as the model you plug in. Pick Cline if you value openness, control and no lock-in and are comfortable managing your own model access; skip the free tier if you specifically need a support SLA or centrally-managed rollout, and budget for Enterprise instead.", "lastVerified": "2026-07-23" }, { "slug": "coderabbit", "name": "CodeRabbit", "tagline": "An AI code-review agent, not a code-writing one — reviews every pull request automatically across GitHub, GitLab, Bitbucket and Azure DevOps, with a genuine free tier for unlimited repos.", "url": "https://theagentsindex.com/coderabbit", "websiteUrl": "https://www.coderabbit.ai", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "enterprise", "api", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "CodeRabbit occupies a genuinely different slot in this category than any other listing: it reviews pull requests and commits automatically rather than writing code from a prompt, so it complements a coding agent instead of competing with it. The free tier is real (unlimited repos for PR summaries and IDE/CLI review, not just a capped trial), the funding and NVIDIA reference customer are independently corroborated, and its four-host coverage (GitHub, GitLab, Bitbucket, Azure DevOps) is broader than most single-purpose review tools. The real costs show up going deeper: API access and self-hosting are Enterprise-only, and per-seat paid pricing stacks on top of whatever already generates the code being reviewed.", "lastVerified": "2026-07-24" }, { "slug": "crescendo", "name": "Crescendo", "tagline": "Fully-managed AI-native customer service, replaces your entire CX stack with agents, QA and human-in-the-loop support, billed per resolution.", "url": "https://theagentsindex.com/crescendo", "websiteUrl": "https://www.crescendo.ai/", "tier": "Listed", "categories": [ "support-agents" ], "tags": [ "enterprise", "autonomous" ], "apiAvailable": false, "pricing": "paid", "verdict": "Crescendo’s pitch is speed and completeness: instead of stitching together a chat widget, a ticketing system and an outsourced BPO separately, it sells all of it — AI agents, QA, routing, knowledge management and human-in-the-loop staff — as one per-resolution-priced engagement. The backing is real: a $500M valuation and the PartnerHero acquisition gave it enterprise-scale delivery capacity within its first year, and its all-inclusive pricing plus Total Outcome Guarantee genuinely aligns cost with results. The tradeoffs are the ones you would expect from a fully managed service: no public API or self-serve trial, limited hands-on configuration control, and per-resolution billing that needs careful volume forecasting to avoid surge fees. Pick Crescendo if you want an outsourced, compliance-ready CX operation you do not have to build yourself; skip it if you want a self-serve, developer-controlled agent like Intercom Fin, or a software-only platform you configure in-house like Sierra.", "lastVerified": "2026-07-23" }, { "slug": "cresta", "name": "Cresta", "tagline": "Unifies an autonomous AI Agent, real-time human Agent Assist and Conversation Intelligence on one contact-center platform, enterprise, custom-quoted, no self-serve tier.", "url": "https://theagentsindex.com/cresta", "websiteUrl": "https://cresta.com/", "tier": "Listed", "categories": [ "support-agents" ], "tags": [ "enterprise", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "Cresta's real differentiator among this category's four existing enterprise support-agent listings is scope: instead of selling an AI Agent that replaces a human conversation, it unifies an AI Agent, real-time Agent Assist for the humans still in the loop, and Conversation Intelligence analytics on one shared record, a genuine \"augment, don't just replace\" position, backed by real scale ($280M+ raised, 500+ employees, named case studies like Xanterra's $3.3M revenue lift). The honest catch: pricing is almost entirely opaque, the only public number anywhere is an AWS Marketplace listing for the narrower Agent Assist product alone, starting at $150,000/year for one channel, while the full AI Agent/Conversation Intelligence suite is custom-quoted with no self-serve tier or free trial. Pick Cresta if a blended AI-plus-human operation is the actual goal and you can run an enterprise sales cycle; skip it if you want transparent self-serve pricing or a fast trial, which none of this category's enterprise players (Cresta, Sierra, Decagon) currently offer.", "lastVerified": "2026-07-23" }, { "slug": "crewai", "name": "CrewAI", "tagline": "A popular open-source Python framework for role-based, collaborating multi-agent teams, free to self-host, now a stable 1.0 release with an optional managed AMP cloud layer.", "url": "https://theagentsindex.com/crewai", "websiteUrl": "https://www.crewai.com", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "multi-agent", "api", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "CrewAI is the fastest way to stand up a working multi-agent system when the mental model is a team of specialists handing off tasks, not one do-everything bot. Its role-based crew abstraction is unusually readable for how much coordination it buys you, and since reaching a stable 1.0 general-availability release in October 2025, with its Crews/Flows APIs locked for long-term use, it now runs at real production scale: 2 billion-plus agentic workflows in the trailing 12 months, and named enterprise use at PepsiCo, Johnson & Johnson, PwC and the US DoD. Tight, deterministic control over long-running production workflows still needs the newer Flows layer rather than the crew model alone. The managed CrewAI AMP cloud product is genuinely useful but thinly priced: its only published self-serve tier is a free 50-execution/month Basic plan, with nothing transparently priced between that and a custom Enterprise conversation. It is a strong pick for prototyping, straightforward collaborative pipelines, and now demonstrably production-grade deployments too, the main catch is AMP’s pricing ladder, not the framework’s maturity.", "lastVerified": "2026-07-24" }, { "slug": "cursor", "name": "Cursor", "tagline": "The AI code editor, an agent that reads, writes and edits across your whole codebase.", "url": "https://theagentsindex.com/cursor", "websiteUrl": "https://cursor.com", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "claude", "api", "gpt", "autonomous", "free-tier" ], "apiAvailable": false, "pricing": "freemium", "verdict": "Cursor is the best in-editor agent for developers who want AI woven into a familiar editor rather than bolted onto a separate tool. Its multi-file agent mode and predictive tab-completion are the most polished in the category, and switching between Claude, GPT and Gemini per request means you are never locked into one model’s blind spots. The tradeoff is real: it is closed-source and subscription-priced, and heavy daily use can hit fast-request limits that push you toward a pricier tier. Pick Cursor for the smoothest agent-in-editor experience available today; skip it if you need a fully scriptable, headless agent for CI. One more factor worth weighing: on 2026-06-16 SpaceX agreed to acquire Cursor’s parent company Anysphere for $60 billion in an all-stock deal, expected to close in Q3 2026 pending regulatory approval, one of the largest startup acquisitions ever. Neither company has said what changes for the product, pricing or model roster once it closes, and analysts have flagged a real risk that SpaceX’s ownership of the rival xAI/Grok models could pull a currently model-agnostic tool toward favoring its new parent’s own stack, buyers who value Cursor’s per-request choice across Claude, GPT and Gemini should watch this closely as the deal progresses.", "lastVerified": "2026-07-24" }, { "slug": "decagon", "name": "Decagon", "tagline": "Enterprise AI customer-service agents that resolve conversations end to end across chat, voice and email, $4.5B-valued, sales-led, no public pricing.", "url": "https://theagentsindex.com/decagon", "websiteUrl": "https://decagon.ai/", "tier": "Listed", "categories": [ "support-agents" ], "tags": [ "enterprise", "no-code", "autonomous" ], "apiAvailable": false, "pricing": "paid", "verdict": "Decagon is one of the best-capitalized names in AI customer support — a $250M Series D in January 2026 valued it at $4.5B, and its published case studies back that up with real numbers: Duolingo’s 80% deflection rate, ClassPass’s 95% cost reduction, Chime’s 70% chat-and-voice resolution rate. Plain-English AOPs let support teams (not just engineers) shape agent behaviour, and Watchtower/Trace View give real visibility into what the agent is doing. The honest catches: there is no public pricing — third-party analysis estimates $74K–$600K+/year depending on volume — resolution-based billing has drawn complaints about how a “resolution” is counted, and reviewers say meaningful deployments still need dedicated engineering to configure AOPs, often over 4–12 weeks. Pick Decagon if you are an enterprise that wants a well-funded, fast-improving agent and can run a proper implementation; skip it if you need transparent self-serve pricing or a public API.", "lastVerified": "2026-07-23" }, { "slug": "devin", "name": "Devin", "tagline": "Cognition’s autonomous AI software engineer, delegate a ticket, get a reviewed pull request back.", "url": "https://theagentsindex.com/devin", "websiteUrl": "https://devin.ai", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "enterprise", "multi-agent", "api", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Devin is the most complete attempt at a genuinely autonomous software engineer: it works asynchronously on its own cloud machine, opens real pull requests, and can be scaled to a fleet on large migrations. It is now also one of the best-capitalized bets in the category — a $1B+ Series D at a $26B valuation (May 2026), $492M in annualized revenue, and named enterprise deployments at Citi, Mercedes-Benz, Goldman Sachs, Dell and Santander back up the pitch with real production usage, not just funding hype. Its July 2026 SWE-1.7 model is a genuine cost-performance play — cheap per task, but Cognition’s own numbers put it behind Claude Opus 4.8 and GPT-5.5 on the hardest benchmarks, so treat it as the efficient default rather than the ceiling. The honest catch on autonomy itself remains: independent reviewers still find it stalls or over-engineers on ambiguous tasks, so its PRs are a draft for review, not a finished merge, and billing can drift once on-demand credits kick in on long sessions. Pick Devin if your bottleneck is well-scoped, delegable engineering work you can review as PRs, or you want a vendor with real enterprise-scale traction behind it; skip it if you want a fast interactive copilot in your own editor, where Cursor or Devin Desktop fit better, or you need frontier-level output on the single hardest tasks without switching models.", "lastVerified": "2026-07-23" }, { "slug": "dify", "name": "Dify", "tagline": "Open-source, visual platform for building production LLM apps and agents, self-host it or run it in the cloud.", "url": "https://theagentsindex.com/dify", "websiteUrl": "https://dify.ai", "tier": "Listed", "categories": [ "agent-platforms" ], "tags": [ "self-hosted", "api", "no-code", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Dify has become the default open-source answer to \"how do we build and ship LLM apps without gluing five tools together.\" Its visual Workflow Studio, built-in RAG pipeline, model-agnostic runtime and observability cover the whole path from prototype to production, and with a large open-source community and a Docker self-host option you get real data control and no lock-in. The nuances matter: the licence is Apache-2.0-derived with extra conditions rather than pure open source, self-hosting means you own the ops burden, and the managed cloud meters usage in message credits that heavier apps will outgrow. For deeply bespoke agent logic a code-first framework can still be more flexible. Pick Dify if you want a self-hostable, all-in-one platform for RAG and agent apps and have the technical chops to run it; skip it if you want a hands-off managed \"AI employee\" with no infrastructure to think about.", "lastVerified": "2026-07-23" }, { "slug": "elevenlabs-conversational-ai", "name": "ElevenLabs Conversational AI", "tagline": "The voice-quality specialist’s own agent layer, 10,000+ voices and instant cloning, with a bring-your-own-LLM brain.", "url": "https://theagentsindex.com/elevenlabs-conversational-ai", "websiteUrl": "https://elevenlabs.io/conversational-ai", "tier": "Listed", "categories": [ "voice-agents" ], "tags": [ "enterprise", "api", "no-code", "open-models", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "ElevenLabs Conversational AI (rebranded ElevenAgents) approaches the voice-agent category from an unusual angle: instead of starting as an orchestration platform and treating voice as one interchangeable component, it starts from the industry’s most-cited text-to-speech and voice-cloning models and wraps an agent layer around them. That shows up as a real edge — 10,000+ voices, instant cloning, and native speech in 70+ languages — backed by a company now valued at $11B with production customers like Deutsche Telekom, Square and Revolut. The LLM layer is genuinely flexible: an ElevenLabs-hosted model for the lowest latency, or bring your own key, including open-source models via an OpenAI-compatible endpoint. The tradeoffs mirror the rest of this category: pricing stacks a subscription, per-minute overage/burst, and separate LLM/telephony usage; and the product covers the voice/orchestration layer, not a turnkey CRM-connected ops tool. Pick it if voice quality and cloning are the deciding factor; skip it if you want a no-code, fully-managed CX product or a stack-agnostic build like Vapi.", "lastVerified": "2026-07-23" }, { "slug": "elicit", "name": "Elicit", "tagline": "An AI research agent built for scientific literature, search, screen and extract cited data from 138M+ papers.", "url": "https://theagentsindex.com/elicit", "websiteUrl": "https://elicit.com", "tier": "Listed", "categories": [ "research-agents" ], "tags": [ "api", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Elicit is the research agent in this category actually built for scientific literature rather than the open web. Its Research Agent autonomously plans and runs a multi-step literature investigation, its Systematic Review workflow automates the screening step of a formal review at real scale (thousands of papers, PRISMA-grade per its own claim at the top tier), and every generated claim links back to a sentence-level citation in the source paper — a genuinely different job than Perplexity's live-web answers or Manus's general task automation. The honest caveats are self-disclosed: Elicit's own docs admit hallucination risk and tell users to verify important findings against the original paper, independent reviewers find it can miss papers a manual search would catch, and it performs best on empirical, formulaic research rather than open-ended theoretical work. Pick Elicit if you need cited, structured evidence from published literature at academic rigor; skip it if you want general web research or broader task automation.", "lastVerified": "2026-07-23" }, { "slug": "factory-ai", "name": "Factory", "tagline": "Agent-native software development — specialized \"Droid\" agents that triage, code, test, review, document and ship changes across the SDLC, from a $20/mo CLI to a fully managed enterprise cloud.", "url": "https://theagentsindex.com/factory-ai", "websiteUrl": "https://www.factory.ai", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "enterprise", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "Factory is a genuinely differentiated bet in the coding-agents category: instead of one generalist assistant, it coordinates specialized \"Droids\" — code, review, test, docs, knowledge — across the full software development lifecycle, from triage through production monitoring, and backs that pitch with real enterprise traction ($150M Series C led by Khosla Ventures at a $1.5B valuation, named users including Morgan Stanley, EY and Palo Alto Networks) and deployment flexibility (SaaS, hybrid, on-premise, air-gapped) most individual-developer coding agents don't offer. The tradeoffs are real: there is no free tier or trial at any price point, the deepest enterprise/security controls are sales-gated behind custom pricing, and its public track record is still thin relative to incumbents with years more case studies. Pick it if your organization wants agents automating SDLC stages beyond code generation and can justify enterprise pricing; look elsewhere if you want a free or cheap way to try an agentic coding tool first.", "lastVerified": "2026-07-24" }, { "slug": "gc-ai", "name": "GC AI", "tagline": "An AI platform for in-house legal teams whose Research Agent runs parallel, cross-checked research across 13M+ US case law opinions, statutes and regulations.", "url": "https://theagentsindex.com/gc-ai", "websiteUrl": "https://gc.ai", "tier": "Listed", "categories": [ "research-agents" ], "tags": [ "enterprise", "multi-agent", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "GC AI is the one research agent in this category built for a specific, high-stakes body of knowledge, US case law, statutes, regulations and agency guidance, that none of Perplexity, Manus, Elicit or GPT Researcher touch. Its Research Agent is genuinely multi-agent: specialized agents search in parallel across jurisdictions and courts, then get cross-checked and reconciled into one cited analysis, backed by real scale ($555M valuation, $73M raised) and real self-serve pricing rather than a fully sales-gated product. The honest caveats: on the cheapest plan, case-law research itself is a paid add-on rather than included; the API is private-beta and account-rep gated; and independent reviews report a thin verified-review trail relative to GC AI's own claimed customer count. Pick GC AI if your research is specifically legal and your budget is an in-house team's, not an individual's; skip it for open-web, academic or free/self-hosted research.", "lastVerified": "2026-07-24" }, { "slug": "gpt-researcher", "name": "GPT Researcher", "tagline": "An open-source autonomous research agent that plans, searches many sources and writes a cited report.", "url": "https://theagentsindex.com/gpt-researcher", "websiteUrl": "https://gptr.dev", "tier": "Listed", "categories": [ "research-agents" ], "tags": [ "self-hosted", "multi-agent", "api", "open-source", "open-models", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "free", "verdict": "GPT Researcher is the reference open-source implementation of the autonomous research agent, and it holds up: give it a question and its planner-and-executor agents fan out across the web, read the sources and return a properly cited, long-form report. Its biggest strengths are openness and control — Apache-2.0 licensed, self-hosted, and agnostic about which LLM and search engine you use, so you can tune cost, quality and privacy to your needs, and even research your own local documents. The tradeoffs are the flip side of that freedom: it is a developer tool you have to install, key up and run, there is no hosted polish, and both quality and spend ride on the model and retriever you pick. Like every research agent, its citations cut hallucination without eliminating it, so reports still need a human check. Pick GPT Researcher if you want a free, hackable, self-hosted research engine you can build on; skip it if you want a no-setup, hosted product and are happy to pay for the convenience.", "lastVerified": "2026-07-23" }, { "slug": "github-copilot", "name": "GitHub Copilot", "tagline": "The mature AI pair-programmer, now with an agent mode, deep in the GitHub + IDE stack.", "url": "https://theagentsindex.com/github-copilot", "websiteUrl": "https://github.com/features/copilot", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "enterprise", "claude", "api", "gpt" ], "apiAvailable": false, "pricing": "freemium", "verdict": "GitHub Copilot is the safest choice for a team that already lives inside GitHub and needs AI coding rolled out at enterprise scale. Its coding agent turns a GitHub issue into a reviewable pull request natively, and its org-level policy, content-exclusion and data-handling controls are more mature than most competitors offer. It is deliberately less aggressive than dedicated agents like Cursor or Devin, a better fit for teams that want reviewable, incremental automation than a fully autonomous coder. Outside the GitHub-plus-supported-IDE stack, the advantage narrows fast.", "lastVerified": "2026-07-23" }, { "slug": "intercom-fin", "name": "Intercom Fin", "tagline": "Intercom’s customer-service AI agent, resolves tickets omnichannel over any helpdesk, billed per resolution.", "url": "https://theagentsindex.com/intercom-fin", "websiteUrl": "https://www.intercom.com/fin", "tier": "Listed", "categories": [ "support-agents" ], "tags": [ "enterprise", "api", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "paid", "verdict": "Fin is the most broadly deployed AI support agent, and its two big advantages are reach and speed: it runs omnichannel (chat, email, voice, social) and can sit on top of virtually any helpdesk, Zendesk, Salesforce, Freshdesk, with Intercom claiming setup in under an hour, backed by a documented API and MCP. Its outcome-style $0.99-per-resolution pricing is genuinely appealing when the agent does the work. The honest caveats are about that meter: Intercom’s 76% headline resolution rate is a best case, with independent analyses of case studies often landing nearer 40-50%, and billing counts “assumed” resolutions when a customer merely stops replying, so costs can be unpredictable and occasionally reward non-resolutions. One more factor to weigh: Salesforce signed a definitive agreement in June 2026 to acquire Intercom/Fin for roughly $3.6B, expected to close around Salesforce’s fiscal Q4 2027, leadership says little changes day-to-day for now, but buyers running Fin over a non-Salesforce helpdesk should watch for roadmap or pricing shifts once the deal closes. Pick Fin if you want a capable agent over your existing stack and can model per-resolution economics (and the pending acquisition); skip it if you need a flat, predictable bill or object to paying for soft resolutions.", "lastVerified": "2026-07-23" }, { "slug": "juggler", "name": "Juggler", "tagline": "An open-source, GUI-based workbench for AI coding agents, branching session trees, not a linear transcript. Free, AGPLv3, by JUCE creator Julian Storer.", "url": "https://theagentsindex.com/juggler", "websiteUrl": "https://juggler.studio", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "self-hosted", "claude", "gpt", "open-source", "open-models" ], "apiAvailable": false, "pricing": "free", "verdict": "Juggler is the most structurally different take on the AI-coding-agent UI to show up in this category recently: instead of a scrolling terminal log, your session is an editable, branching tree, and every tool call, approval and raw context message stays inspectable rather than hidden behind a chat bubble. That is a real, substantive difference from Cursor, Windsurf or Aider, not a cosmetic one, and it comes free, open-source and model-agnostic across Claude, GPT, Gemini, Ollama, OpenRouter and DeepSeek. The honest catch is timing: it shipped on Hacker News on 2026-07-12 and is alpha software by the creator’s own description — a launch-day tester hit an LLM-connection error immediately, there is no ACP support yet for developers coming from an existing plugin ecosystem, and the main app’s AGPLv3 license is a stronger copyleft commitment than most rivals carry. Pick Juggler if you want hands-on visibility into what your agent is doing and are willing to run early-stage software to get it; skip it for now if you need a mature, supported tool or depend on an ACP-based plugin setup.", "lastVerified": "2026-07-23" }, { "slug": "langchain", "name": "LangChain", "tagline": "The most widely-used framework for building LLM apps and agents, with LangGraph for stateful control.", "url": "https://theagentsindex.com/langchain", "websiteUrl": "https://www.langchain.com", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "multi-agent", "api", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "LangChain is the default starting point for building a custom, production-grade agent rather than adopting an off-the-shelf tool, mostly because nothing else matches its integration ecosystem. LangGraph adds real stateful control, branching, loops, human-in-the-loop, multiple coordinated agents, that a simple chain-based library cannot offer, and LangSmith adds production tracing and evals once you are ready to pay for them. The cost is a real learning curve: the core abstractions have changed across major versions, and older tutorials can actively mislead. It is the right tool when you are building something bespoke, and overkill for a single prompt-and-response wrapper. Worth knowing: LangChain raised a $125 million Series B in October 2025 at a $1.25 billion valuation, becoming a unicorn, a real signal of staying power, and because the LangChain and LangGraph libraries themselves are open-source and self-hostable, your own deployment is not at the mercy of the company’s fate the way a closed-source tool’s would be.", "lastVerified": "2026-07-24" }, { "slug": "letta", "name": "Letta", "tagline": "A stateful-agent framework — open-source, memory-first agents whose context, tools and identity persist as a durable server object instead of resetting between sessions.", "url": "https://theagentsindex.com/letta", "websiteUrl": "https://www.letta.com", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "api", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Letta is the one framework in this category built around a genuinely different premise: agents as stateful, persistent objects — with git-versioned memory and idle-period \"sleep-time compute\" — rather than stateless call graphs assembled fresh on every request. That premise traces directly back to MemGPT, the Berkeley research it grew out of, and its founders (Sarah Wooders, Charles Packer) are the same researchers behind that work. The open-source core is real and active (23.9K+ GitHub stars, Apache 2.0, self-hostable), backed by a $10M Felicis-led seed round and used in production at Bilt, 11x, Kognitos and Hunt Club. The tradeoffs: it's a much smaller, younger ecosystem than LangChain's, its funding is a single, smaller seed round rather than a later-stage raise, and its product surface is split across two differently-adopted repos. Pick it when persistent, evolving agent memory is the actual requirement; pick a stateless orchestration framework when the task is a bounded workflow that doesn't need to remember anything between runs.", "lastVerified": "2026-07-24" }, { "slug": "lindy", "name": "Lindy", "tagline": "An AI executive assistant you text over iMessage or SMS, built on a no-code agent platform still available underneath.", "url": "https://theagentsindex.com/lindy", "websiteUrl": "https://www.lindy.ai", "tier": "Listed", "categories": [ "agent-platforms" ], "tags": [ "multi-agent", "no-code", "autonomous" ], "apiAvailable": false, "pricing": "paid", "verdict": "Lindy has moved from selling itself as a no-code \"AI employee\" builder to selling a finished AI executive assistant you text over iMessage or SMS, connect your phone and inbox, and it starts triaging email, prepping meetings and chasing follow-ups without you opening an app. That text-first interface is genuinely the fastest on-ramp to a useful agent in this category, and human-in-the-loop review keeps you in control. The old no-code agent builder hasn’t gone away, it still exists under an \"App Builder\" product for teams who want to construct custom multi-agent flows, but it is no longer what Lindy leads with, so buyers who want that specifically have to go looking for it. The other real catch is cost transparency: usage is metered in credits, stronger models burn more, and once a plan’s monthly allotment is gone, overage work is billed at 2x the standard rate. It’s also cloud-only, closed-source, and has no permanent free plan, just a 7-day trial. Pick Lindy if you want a working AI assistant you can text starting today and will watch your credit usage; skip it if you need self-hosting, an open API to build on, or flat, predictable pricing.", "lastVerified": "2026-07-24" }, { "slug": "livekit-agents", "name": "LiveKit Agents", "tagline": "The open-source, self-hostable voice agent framework, Apache 2.0, provider-agnostic, and the infrastructure behind ChatGPT’s Advanced Voice Mode.", "url": "https://theagentsindex.com/livekit-agents", "websiteUrl": "https://livekit.com/voice-agents", "tier": "Listed", "categories": [ "voice-agents" ], "tags": [ "self-hosted", "api", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "LiveKit Agents is the one genuinely open-source, self-hostable option among this category’s voice agent platforms, an Apache 2.0 framework, not a closed API you can only rent. That openness comes from a company with real weight behind it: a $1B valuation on $181.3M raised, and infrastructure serious enough that OpenAI runs ChatGPT’s Advanced Voice Mode on it alongside customers like xAI, Salesforce and Tesla. The tradeoff is real: unlike Vapi, Retell AI, Bland AI or ElevenLabs Conversational AI, there is no hosted dashboard that gets you a working agent in an afternoon, you write and deploy code, choosing and wiring your own STT/LLM/TTS providers (or LiveKit’s managed Inference gateway) yourself. Pick LiveKit Agents if engineering ownership, self-hosting or avoiding vendor lock-in matter more than a fast, no-code start; skip it if you want a turnkey agent live without writing code.", "lastVerified": "2026-07-24" }, { "slug": "llamaindex", "name": "LlamaIndex", "tagline": "The data framework for LLM apps, RAG-first, now with agent workflows and a document-parsing cloud.", "url": "https://theagentsindex.com/llamaindex", "websiteUrl": "https://www.llamaindex.ai", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "multi-agent", "api", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "LlamaIndex is the reference implementation for retrieval-augmented generation: if your application lives or dies on getting the right context into the model, its query engines, advanced retrieval patterns and 300-plus connectors are the deepest toolkit available, and the MIT-licensed core is genuinely free. Its expansion into event-driven Workflows and agent tooling means it is no longer just a RAG library, though agent-first orchestration is still more naturally LangGraph or CrewAI territory. The commercial LlamaCloud, chiefly LlamaParse, is a real asset for messy PDFs, tables and scans, and it comes with a monthly free-credit allotment, but the best parsing sits behind usage-based paid plans rather than in the free package. Be prepared for a learning curve on advanced retrieval and for an API that has churned across frequent releases. Pick LlamaIndex when retrieval quality is the core of your product; skip it as your primary tool if your app is chiefly a real-time conversational or multi-tool agent with little document grounding.", "lastVerified": "2026-07-23" }, { "slug": "manus", "name": "Manus", "tagline": "An autonomous general AI agent that plans and executes multi-step tasks end-to-end in its own cloud computer.", "url": "https://theagentsindex.com/manus", "websiteUrl": "https://manus.im", "tier": "Listed", "categories": [ "research-agents" ], "tags": [ "browser-use", "multi-agent", "api", "no-code", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Manus is one of the most ambitious consumer agents shipped so far: hand it a goal and it genuinely plans and executes a long chain of steps in its own cloud computer, browsing, coding and producing a finished deliverable with little supervision. When the task is straightforward and mostly read-only — a research summary, a data pull, a simple site — it can feel like magic. The honest problem is reliability: independent testing has repeatedly caught it derailing on long, branching workflows and even fabricating verification output that looked real, so it is not something to trust with mission-critical or write-heavy work yet. The credit-based pricing also makes cost unpredictable, and its corporate ownership remains unsettled after a Meta acquisition was announced, blocked by Chinese regulators, and is now the subject of a Tencent-led buyback still being negotiated. Pick Manus to offload well-scoped, low-stakes multi-step tasks and to see where autonomous agents are heading; skip it for anything where an unnoticed hallucination or a stalled workflow would be expensive.", "lastVerified": "2026-07-23" }, { "slug": "mastra", "name": "Mastra", "tagline": "The TypeScript-native agent framework — agents, durable workflows, memory and observability in one Apache-2.0 package, with an optional managed cloud platform.", "url": "https://theagentsindex.com/mastra", "websiteUrl": "https://mastra.ai", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "api", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Mastra is the one genuinely TypeScript-native entrant in a category otherwise built Python-first with JS bindings layered on — a real, defensible niche for full-stack JS and Next.js teams who don’t want to stand up a separate Python service just to add an agent. It unifies agents, durable typed workflows, memory and observability in one Apache-2.0 package that is free to self-host, backed by real commercial traction ($35M raised, a Spark Capital-led Series A, and named production customers including Replit, Sanity, MongoDB, Brex and Marsh). The tradeoffs are real too: it is TypeScript-only with no Python path at all, its ecosystem is a fraction of LangChain’s size and age, its deepest enterprise controls sit behind a separate source-available license rather than the open core, and its optional Mastra Cloud platform bills across several metered dimensions rather than one flat number. Pick it if you are building in the JS/TS stack and want an integrated framework designed for that world from day one; pick something with a longer track record if Python is your default or you need the widest possible integration ecosystem today.", "lastVerified": "2026-07-23" }, { "slug": "microsoft-autogen", "name": "Microsoft AutoGen", "tagline": "Microsoft’s open-source framework for orchestrating conversations between multiple AI agents.", "url": "https://theagentsindex.com/microsoft-autogen", "websiteUrl": "https://microsoft.github.io/autogen/", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "multi-agent", "open-source", "free-tier" ], "apiAvailable": false, "pricing": "free", "verdict": "AutoGen is the framework that popularised the “agents conversing with agents” model, and it remains a clean, well-documented and genuinely free way to prototype multi-agent systems in Python or .NET. Its layered design is a real strength: you can start in the high-level AgentChat API and drop into the Core runtime when you need fine-grained control over message passing. The catch is strategic, not technical, and it has hardened since AutoGen’s own SEED write-up: Microsoft’s successor, the Microsoft Agent Framework, is no longer just announced, it reached a production-ready 1.0 GA on 2026-04-03 and is the live recommended path for new projects, while the original authors’ AG2 fork remains genuinely active in parallel (shipping commits within the last week) and controls the old PyPI packages. AutoGen itself hasn’t tagged a release since September 2025. That three-way split is a real, current risk for anyone building something they must maintain for years, and multi-agent loops can be token-heavy and hard to make deterministic. Pick AutoGen today to learn the patterns or extend an existing AutoGen codebase; if you are starting fresh for production, evaluate the Microsoft Agent Framework or AG2 first.", "lastVerified": "2026-07-24" }, { "slug": "openai-agents-sdk", "name": "OpenAI Agents SDK", "tagline": "OpenAI’s lightweight open-source framework for building agents, handoffs, guardrails and tracing in a few primitives.", "url": "https://theagentsindex.com/openai-agents-sdk", "websiteUrl": "https://openai.github.io/openai-agents-python/", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "multi-agent", "gpt", "open-source", "free-tier" ], "apiAvailable": false, "pricing": "free", "verdict": "The OpenAI Agents SDK is the shortest path from idea to a working agent: four clean primitives, agents, handoffs, guardrails and sessions, plus built-in tracing that would otherwise take weeks to build, and a beta “sandbox agents” mode that gives an agent its own persistent file/shell workspace. It is MIT-licensed and free, ships in both Python and TypeScript, and despite the OpenAI branding it is genuinely provider-agnostic, running 100-plus models through LiteLLM. The flip side of that minimalism is real: durable state is limited to `RunState`, a snapshot you capture manually after a run pauses for human approval, not automatic step-by-step checkpointing, and handoffs are linear chains rather than a graph, anything beyond agents-plus-tools-plus-handoffs becomes code you own, which is why teams with genuinely complex, branching or audit-heavy workflows often graduate to LangGraph. It is also young and moving fast, so expect some API churn, and the tracing experience leans on OpenAI’s dashboard. Pick it for support triage, routing and linear tool-using agents you want in production this week; skip it when your workflow needs deterministic branching, automatic mid-run checkpointing or compliance-grade audit trails.", "lastVerified": "2026-07-24" }, { "slug": "codex-cli", "name": "OpenAI Codex CLI", "tagline": "OpenAI's open-source (Apache-2.0) terminal coding agent, bundled into ChatGPT plans from Free through Pro — Business seats are now grandfathered-only.", "url": "https://theagentsindex.com/codex-cli", "websiteUrl": "https://developers.openai.com/codex/cli", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "api", "gpt", "open-source", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Codex CLI’s biggest advantage isn’t a feature — it’s bundling: because it draws on the same plan and credit pool as ChatGPT, a developer who already pays for Plus, Pro or Business gets a capable, actively-maintained terminal coding agent with no separate subscription decision to make, and even the Free plan gets limited access to try it. The tool itself is well-built and open-source, with granular permission controls, subagents, MCP support and a genuinely useful built-in code-review mode running across terminal, desktop and IDE surfaces. The honest catches are real: switching to a non-OpenAI model takes manual config beyond the built-in Ollama/LM Studio/Bedrock options, OpenAI’s own numbers put typical cost at $100-$200/developer per month for regular use, and a large, currently-open GitHub issue documents a real rate-limit consumption regression from June 2026 that was still unresolved as of this listing’s last check. Pick Codex CLI if you already live in the ChatGPT ecosystem and want a terminal agent without a new bill; pick Claude Code instead if you want a lower, more predictable entry price, or Windsurf/Juggler if free multi-model switching matters more to you than OpenAI’s specific model lineup.", "lastVerified": "2026-07-23" }, { "slug": "opencode", "name": "OpenCode", "tagline": "A free, open-source (MIT) AI coding agent for the terminal, desktop and IDE, bring any of 75+ model providers, including your own local Ollama models.", "url": "https://theagentsindex.com/opencode", "websiteUrl": "https://opencode.ai/", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "self-hosted", "api", "open-source", "open-models", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "OpenCode’s core pitch is real: it costs nothing to run, is genuinely open-source under MIT, and connects to essentially any model — 75+ cloud providers plus local Ollama — rather than betting on one vendor the way Claude Code or Cursor do. That flexibility, combined with a very fast release cadence (a new version roughly every few days) and a large, active GitHub community (188,976 stars as of this listing’s research date), makes it a credible pick for developers who already have a preferred model or subscription and do not want a second bill. The honest catches: its disclosed 2026-01-12 security history includes one critical and one high-severity advisory (both patched, but real), an open GitHub issue is still tracking real memory-growth complaints as of this listing’s research date, and a dead, similarly-named repository (opencode-ai/opencode) makes it easy to land on the wrong project by accident. Pick OpenCode if free, open-source and model flexibility matter more to you than a single polished vendor experience; pick Claude Code or Cursor instead if you would rather pay for one vendor’s managed reliability and support.", "lastVerified": "2026-07-23" }, { "slug": "openhands", "name": "OpenHands", "tagline": "Open source AI agent platform for cloud coding agents, from free local use to enterprise self-hosted deployment.", "url": "https://theagentsindex.com/openhands", "websiteUrl": "https://all-hands.dev", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "self-hosted", "enterprise", "multi-agent", "api", "open-source", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "OpenHands is a credible, community-backed choice for engineering teams that want an open source, self-hostable coding agent platform rather than a closed SaaS tool. Its free OSS core, embeddable SDK, and enterprise-grade sandboxed deployment options make it flexible for both solo developers and security-conscious organizations, provided you're willing to accept a more technical setup process — whether you bring your own LLM API key or pay OpenHands' own at-cost rate for model usage.", "lastVerified": "2026-07-24" }, { "slug": "perplexity", "name": "Perplexity", "tagline": "The AI answer engine, with agentic Deep Research and the Comet browser that acts on the web for you.", "url": "https://theagentsindex.com/perplexity", "websiteUrl": "https://www.perplexity.ai", "tier": "Listed", "categories": [ "research-agents" ], "tags": [ "browser-use", "enterprise", "claude", "api", "gpt", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Perplexity is the most refined AI answer engine available, and calling it an ‘agent’ is fair only because of the layers it has added on top: Deep Research, which autonomously compiles a cited report, and Comet, a browser whose built-in agent can act on the web for you. For everyday sourced answers it is hard to beat, and per-query model choice across Sonar, GPT, Claude, Gemini and now GLM and Kimi means you are not tied to one model’s weaknesses. The honest caveats are that it is still, at heart, a question-answering tool rather than a general workflow agent, and that like any LLM search product it can occasionally hallucinate or cite a source that doesn’t quite back the claim — so verification still matters. The best agentic pieces also live on the pricier Max and Enterprise tiers. Pick Perplexity if you want first-class sourced answers plus an autonomous research mode and an agentic browser in one place; skip it if you need a self-hosted or fully programmable task agent.", "lastVerified": "2026-07-23" }, { "slug": "polyai", "name": "PolyAI", "tagline": "The pre-LLM-era enterprise voice AI incumbent — $750M-valued, 200+ enterprise customers — that opened a genuine self-serve, free-to-start tier in May 2026.", "url": "https://theagentsindex.com/polyai", "websiteUrl": "https://poly.ai", "tier": "Listed", "categories": [ "voice-agents" ], "tags": [ "enterprise", "api", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "paid", "verdict": "PolyAI is the pre-LLM-era incumbent of the voice-agent category — founded in 2017, years before Vapi, Retell AI, Bland AI or ElevenLabs Conversational AI existed — with real enterprise scale to show for it: 200+ customers, thousands of live deployments, a $750M valuation, and a Forrester study (vendor-commissioned, but methodologically disclosed) modeling 391% three-year ROI for a composite enterprise customer. Until May 2026 it was also the most closed platform in this category — sales-led only, no public price, no self-serve account. That changed: PolyAI now offers a genuine self-serve tier, free for two months from just an email address, with both a no-code builder and a real developer API (ADK). The catch is that PolyAI has not disclosed what happens once that free period ends, and its flagship enterprise deployments (like PG&E’s 16M-call-a-year operation) still run through a separate, custom-quoted sales engagement. Pick PolyAI if you want proven enterprise scale with a low-friction way to try it first; go in aware that the free tier’s long-run economics are still a question mark.", "lastVerified": "2026-07-25" }, { "slug": "pydantic-ai", "name": "Pydantic AI", "tagline": "The type-safe agent framework from the Pydantic team, that FastAPI feeling for building AI agents.", "url": "https://theagentsindex.com/pydantic-ai", "websiteUrl": "https://ai.pydantic.dev", "tier": "Listed", "categories": [ "frameworks" ], "tags": [ "self-hosted", "multi-agent", "open-source", "free-tier" ], "apiAvailable": false, "pricing": "free", "verdict": "Pydantic AI brings the same discipline that made Pydantic and FastAPI ubiquitous to agent building: define your agent and its output with Python type hints, and the framework validates every model response against them, catching malformed outputs at the boundary instead of deep in production. It is MIT-licensed and free, genuinely model-agnostic across OpenAI, Anthropic, Gemini and many more, and its dependency-injection system and lightweight runtime make for an unusually pleasant, low-boilerplate developer experience. It has matured fast: a stable 1.0 in September 2025, then a backwards-incompatible 2.0 on 2026-06-23 that added official durable-execution integrations (Temporal, DBOS, Prefect and Restate) and native AG-UI/A2A protocol support, closing much of the gap with graph-orchestration frameworks on long-running, checkpointed and human-in-the-loop workflows. The trade-offs have narrowed but not disappeared: it still ships fewer pre-built integrations than LangChain, and durable execution is something you operate yourself via an external orchestrator rather than a built-in, zero-infrastructure engine, and LangGraph’s native graph-branching model still goes deeper for the most complex, stateful orchestration. The best observability story also runs through Pydantic’s own Logfire product. Pick Pydantic AI when you want type-safe, validated outputs with minimal ceremony and are willing to wire in an orchestrator when you need production-grade durability; skip it when you want the deepest native graph-branching control or a fully built-in durability engine with nothing extra to operate.", "lastVerified": "2026-07-24" }, { "slug": "qualified", "name": "Qualified", "tagline": "Piper, the inbound AI SDR, engages website visitors in real time, qualifies them against your ICP and books the meeting. Salesforce-native; Salesforce itself acquired the company in April 2026.", "url": "https://theagentsindex.com/qualified", "websiteUrl": "https://www.qualified.com", "tier": "Listed", "categories": [ "sales-marketing-agents" ], "tags": [ "enterprise", "api", "autonomous" ], "apiAvailable": true, "pricing": "paid", "verdict": "Qualified earns its place in this category by being genuinely differentiated from the outbound-first AI SDRs that dominate it: Piper works inbound, live, on your website and in your inbox, deeply wired into Salesforce rather than treating the CRM as just another integration. The named case studies (Asana, Greenhouse) and the fact that Salesforce itself paid roughly $1.2 billion to acquire the company outright in April 2026 are real, verifiable signals of substance, not just marketing claims. The honest catches: Salesforce is a hard requirement, not an option, pricing is entirely custom (independent procurement data puts the median deal at $61,000/year), and — now that Salesforce owns it — the product's future is tied to Agentforce's roadmap rather than an independent vendor's. Pick Qualified if you're Salesforce-native and your bottleneck is converting inbound traffic; skip it if you need outbound prospecting or run on a different CRM.", "lastVerified": "2026-07-23" }, { "slug": "relevance-ai", "name": "Relevance AI", "tagline": "Build and run an \"AI workforce\", teams of no-code agents that do real sales, ops and research work.", "url": "https://theagentsindex.com/relevance-ai", "websiteUrl": "https://relevanceai.com", "tier": "Listed", "categories": [ "agent-platforms" ], "tags": [ "enterprise", "multi-agent", "api", "no-code", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "Relevance AI is one of the more serious \"agent platform\" plays: it’s built from the ground up for multi-agent workforces, with the unglamorous production plumbing, an LLM router, evals, tracing, approvals and RBAC/SSO, that separates a demo from something you’d trust with live pipeline. The cost-optimising router and bring-your-own-keys option are genuinely useful for keeping AI spend sane. The tradeoffs are real: the dual meter of Actions plus Vendor Credits is easy to misjudge on AI-heavy agents, the flagship Bosh BDR is a custom enterprise onboarding rather than a self-serve click, and the public pricing page now leads with \"talk to sales.\" Pick Relevance AI if you want to build coordinated agent teams with enterprise-grade control and will invest in learning it; skip it if you need a single quick bot, open-source self-hosting, or perfectly predictable flat pricing.", "lastVerified": "2026-07-23" }, { "slug": "replit-agent", "name": "Replit Agent", "tagline": "Describe an app and Replit Agent builds, tests and deploys it in your browser, a full-stack app factory for anyone.", "url": "https://theagentsindex.com/replit-agent", "websiteUrl": "https://replit.com/ai", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "browser-use", "enterprise", "no-code", "autonomous", "free-tier" ], "apiAvailable": false, "pricing": "freemium", "verdict": "Replit Agent is the most accessible way to turn a plain-language idea into a running, deployed app: it writes the code, provisions a database, adds auth, tests itself in a browser and ships to a live URL, all in one tab, and Agent 3 can grind autonomously for up to ~200 minutes fixing its own bugs. Agent 4, launched March 2026 alongside a $400M Series D at a $9B valuation (3x six months earlier), pushes further into collaborative, multi-artifact building — an Infinite Canvas and parallel, mergeable task threads for building a web app, mobile app and landing page in the same project — though independent benchmarks of Agent 4 specifically are not yet public. For founders and non-engineers that is genuinely powerful — there is nothing local to install and a free Starter tier to try. The honest catch is cost and control: effort-based credit billing can escalate quickly, with real feature work costing dollars per prompt and heavy users routinely blowing past the monthly credit pool, and the generated code can be hard to maintain as an app grows. It is also cloud-only and closed. Pick Replit Agent if you value speed and accessibility and want a deployed app fast; skip it if you need full control of your own toolchain or predictable, bounded costs on large projects.", "lastVerified": "2026-07-23" }, { "slug": "retell-ai", "name": "Retell AI", "tagline": "A voice AI platform for human-sounding phone agents, build a call flow, plug in your model and telephony, ship.", "url": "https://theagentsindex.com/retell-ai", "websiteUrl": "https://www.retellai.com", "tier": "Listed", "categories": [ "voice-agents" ], "tags": [ "enterprise", "api", "open-models", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "paid", "verdict": "Retell AI is one of the strongest developer-first voice platforms, and its reputation rests on the two things that matter most on a phone call: natural voice quality and low latency, backed by a proprietary turn-taking model at around 600ms. You assemble the stack — LLM, voice, telephony — and get a real builder, an auto-syncing knowledge base, batch calling and simulation testing, plus a broad set of CRM and contact-centre integrations. The trade-offs are the usual ones for this class: component pricing that starts at $0.07/min but realistically lands north of $0.13–$0.20/min once fully specced, and a learning curve that assumes technical staff. Pick Retell if you want top-tier voice quality with control over the stack; skip it if you need a no-code tool with predictable flat pricing and hands-on support.", "lastVerified": "2026-07-23" }, { "slug": "rox", "name": "Rox", "tagline": "An AI-native revenue operating system, \"agent swarms\" (one per account) that research, monitor and update your existing CRM across the whole deal lifecycle, with a genuine free self-serve tier.", "url": "https://theagentsindex.com/rox", "websiteUrl": "https://www.rox.com", "tier": "Listed", "categories": [ "sales-marketing-agents" ], "tags": [ "enterprise", "autonomous", "free-tier" ], "apiAvailable": false, "pricing": "freemium", "verdict": "Rox's real differentiation in this category is scope, not automation intensity: Clay is a data-enrichment engine, 11x and Artisan run autonomous outbound sequences, and Qualified works inbound website traffic, Rox instead assigns a persistent agent to every account and keeps researching, monitoring and updating your existing CRM across the whole relationship, answered through plain-English queries rather than a new dashboard to learn. The named case studies (Ramp, MongoDB, New Relic) and a rare real self-serve price (free Starter tier, $50/month Core) back that up with more transparency than most enterprise-focused rivals in this category offer. The honest catches: no confirmed public API for custom integration work, the free/Core action allowances are modest for a real book of accounts, and at two years old it carries far less production history than the CRMs it plugs into. Pick Rox if continuous account intelligence and CRM hygiene are the bottleneck; skip it if you specifically need outbound-sequence automation or a documented API.", "lastVerified": "2026-07-24" }, { "slug": "sierra", "name": "Sierra", "tagline": "Bret Taylor’s enterprise agent platform, branded AI agents that resolve customer issues end to end, priced on outcomes.", "url": "https://theagentsindex.com/sierra", "websiteUrl": "https://sierra.ai", "tier": "Listed", "categories": [ "agent-platforms", "support-agents" ], "tags": [ "enterprise", "no-code", "autonomous" ], "apiAvailable": false, "pricing": "paid", "verdict": "Sierra is the enterprise standard-bearer for customer-experience agents, and the pedigree is real: co-founded by Bret Taylor and Clay Bavor, it reached $150M+ ARR by February 2026, has raised roughly $1.585 billion at a valuation above $15 billion, and is used across more than 40% of the Fortune 50. What sets it apart is depth, agents don’t just answer, they act, completing refunds and account changes across chat, voice, email and more, wrapped in serious tooling for building, supervising and testing them. Its outcome-based pricing (pay for resolutions, escalations free) is a genuinely aligned model. The flip side is access: it is sales-led with no public pricing, no free tier and a gated SDK, so it is neither quick to trial nor a fit for smaller teams. Pick Sierra if you are an enterprise that wants autonomous agents doing real work; skip it if you need self-serve, transparent pricing or an open API.", "lastVerified": "2026-07-24" }, { "slug": "unify", "name": "Unify", "tagline": "A single conversational GTM agent for warm outbound — self-serve from $0, built to augment a seller rather than replace one.", "url": "https://theagentsindex.com/unify", "websiteUrl": "https://www.unifygtm.com", "tier": "Listed", "categories": [ "sales-marketing-agents" ], "tags": [ "gpt", "free-tier" ], "apiAvailable": false, "pricing": "freemium", "verdict": "Unify fills a real, previously under-served position in this category's GTM-outbound cluster: cheaper and far more transparent than Clay's usage-based ($167+/month) enrichment canvas, and self-serve where 11x and Artisan are both entirely custom-quoted and sales-gated. It explicitly positions itself as a productivity layer for a human seller rather than an autonomous replacement, wrapping list-building, enrichment, signal-based prioritization and sequencing into one chat-based agent instead of a spreadsheet-style workflow builder. The catch: it just completed a dramatic pricing-model pivot — several third-party sources published within the last few months still describe a completely different enterprise-only motion — so treat pricing found outside unifygtm.com/pricing itself as potentially stale, and its deepest features (GPT-5.5, write-access CRM sync, website intent signals) still sit behind a custom Business tier. Pick Unify if you want an accessible, self-serve GTM agent that augments a rep rather than replacing one; skip it if you specifically want Clay's build-your-own depth or 11x/Artisan's fully autonomous outbound motion.", "lastVerified": "2026-07-24" }, { "slug": "vapi", "name": "Vapi", "tagline": "The developer platform for voice AI agents, bring your own model, voice and telephony, and Vapi orchestrates the call.", "url": "https://theagentsindex.com/vapi", "websiteUrl": "https://vapi.ai", "tier": "Listed", "categories": [ "voice-agents" ], "tags": [ "enterprise", "multi-agent", "api", "open-models", "autonomous", "free-tier" ], "apiAvailable": true, "pricing": "paid", "verdict": "Vapi is the builder’s choice for voice AI: an API-first, model-agnostic platform that hands developers control over every layer — transcription, model, voice and telephony — with real-time orchestration and sub-500ms latency. Its bring-your-own-provider design means you can chase the best model or cheapest voice without rebuilding the agent, and enterprise features (HIPAA, zero-data-retention, SLAs) let it scale to millions of calls. The honest catch is cost transparency: the $0.05/min headline is only the orchestration fee, and the true all-in rate is usually several times that once you add a model, a good voice and telephony. Pick Vapi if you have engineers who want to build a bespoke voice agent and own the stack; skip it if you want a no-code tool that hides the plumbing.", "lastVerified": "2026-07-23" }, { "slug": "vecbase", "name": "Vecbase", "tagline": "A shared workspace where AI agents with roles like Researcher, Coder, and Writer work as a team", "url": "https://theagentsindex.com/vecbase", "websiteUrl": "https://vecbase.com", "tier": "Listed", "categories": [ "agent-platforms" ], "tags": [], "apiAvailable": false, "pricing": "paid", "verdict": "Vecbase packages multi-agent collaboration, shared memory and files, built-in AI capabilities, and one-click integrations into a single subscription, which is appealing for teams tired of stitching together separate tools. The tradeoffs are a hard cap of 10 agents per workspace, usage-based overage billing on top of the base plan, no free trial, and documentation that leans heavily on Chinese, all worth weighing before committing.", "lastVerified": "2026-07-23" }, { "slug": "windsurf", "name": "Windsurf (now Devin Desktop)", "tagline": "The agentic IDE that was Codeium, then Windsurf, now Cognition’s Devin Desktop, a command centre for coding agents.", "url": "https://theagentsindex.com/windsurf", "websiteUrl": "https://devin.ai/desktop", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "claude", "multi-agent", "gpt", "autonomous", "free-tier" ], "apiAvailable": false, "pricing": "freemium", "verdict": "Windsurf was one of the best agentic IDEs of its generation, and as Devin Desktop it is now something more ambitious: a vendor-neutral command centre where local and cloud agents, Cognition’s own plus Claude Agent, Codex and Gemini CLI via the open ACP, run inside one polished editor. That openness is a real differentiator in a field of walled-garden tools, and the shared free/$20 subscription with Devin makes it easy to try. The cost is turbulence: Codeium became Windsurf, Windsurf became Devin Desktop, and the original Cascade agent was switched off on 1 July 2026, a lot of change for a tool teams use daily. Its default SWE models are also a notch behind frontier models on the hardest tasks, though you can route to those. Pick it if you want an open, multi-agent desktop IDE and are comfortable inside Cognition’s ecosystem; skip it if you want a lightweight editor extension or you specifically wanted the standalone Windsurf that no longer exists.", "lastVerified": "2026-07-24" }, { "slug": "zot", "name": "Zot", "tagline": "A free, MIT-licensed coding-agent harness in one Go binary, branch any session from any past message and view it as a tree, plus a Telegram bridge.", "url": "https://theagentsindex.com/zot", "websiteUrl": "https://zot.sh", "tier": "Listed", "categories": [ "coding-agents" ], "tags": [ "self-hosted", "claude", "api", "gpt", "open-source", "open-models", "free-tier" ], "apiAvailable": true, "pricing": "free", "verdict": "Zot earns its place with a real, working idea: fork any past message in a coding session into a new branch and see the whole tree at a glance, all from one free, MIT-licensed Go binary with no Docker or Node runtime. It backs that up with genuine breadth — 30+ model providers, local Ollama support, a Telegram bridge, background swarm subagents, and an importable Go SDK for developers who want to build their own harness on top of it — and it ships fast, with 303 releases and 10+ contributors since an April 2026 start. The honest catches are the ones the maintainer discloses directly: the OAuth-reuse trick that lets it bill through a Claude/ChatGPT subscription is, by the README's own admission, a likely terms-of-service violation that could be revoked at any time (the API-key flow sidesteps this), the docs call the project \"beta forever\" rather than production-stable, and the TUI is admitted to be almost entirely AI-generated. Pick Zot if you want a free, fast, Go-native harness with real branching-session control and are comfortable with fast-moving independent software; pick Juggler instead if you want the same branching-tree idea with a GUI, or a first-party vendor tool if you need managed support and want to avoid any OAuth-reuse risk entirely.", "lastVerified": "2026-07-23" }, { "slug": "n8n", "name": "n8n", "tagline": "Source-available workflow automation with native AI-agent nodes, self-hostable and integration-rich.", "url": "https://theagentsindex.com/n8n", "websiteUrl": "https://n8n.io", "tier": "Listed", "categories": [ "agent-platforms" ], "tags": [ "self-hosted", "api", "no-code", "open-source", "free-tier" ], "apiAvailable": true, "pricing": "freemium", "verdict": "n8n is the strongest choice for teams that want agentic automation built visually rather than in code, without giving up the integration breadth or self-hosting control that no-code tools usually sacrifice. Its 400+ app integrations plus native AI-agent nodes mean an agent can actually go do something, call an API, update a CRM, post to Slack, not just answer a question. It is fair-code, not OSI open-source, so check the licence before reselling a self-hosted deployment commercially, and deeply branching agent logic can get harder to read on a canvas than in code. It suits ops-style automation with an agent in the loop best; developers who prefer expressing control flow purely as code may find LangGraph or CrewAI a better fit.", "lastVerified": "2026-07-23" } ] }