# Agents·Index

> AI agents, frameworks and agentic tools you can actually get: what each one
> does, what it costs and whether it ships. Every listing sourced, checked and
> dated.

The Agents Index is a curated, quality-gated directory of agents. Every
listing is a researched record that must pass an automated quality gate —
structured fields, sourced facts, an on-topic fit-check — before it can
publish. Thin, auto-generated pages are structurally impossible, and every
listing carries a last-verified date.

## The directory right now

- 146 quality-gated agents published.
- Freshest verification: 2026-10-09.

## Machine-readable surfaces (read + cite these)

- https://theagentsindex.com/listings.json — the full Published set as JSON (structured facts per agent: name, url, tier, categories, tags, pricing, API, last-verified). The machine-readable source of truth.
- https://theagentsindex.com/rss.xml — RSS feed of listings, newest-verified first.
- https://theagentsindex.com/index.md — the home twin: what this index is, the live counts, every category with its published count, the picks and the most-recently re-verified rows. Start here for the entity, then follow a slug.
- https://theagentsindex.com/llms-full.txt — the WHOLE corpus as one markdown document (every listing, answer-first) for one-shot ingestion.
- Per-listing markdown twins at https://theagentsindex.com/{slug}.md — the same researched facts as the HTML page, token-lean.
- https://theagentsindex.com/mcp — the MCP door (JSON-RPC POST): search_listings / get_listing / list_categories / compare_listings / get_best_of (free, keyless reads) + submit_listing (the write door). Full tool schemas: https://theagentsindex.com/.well-known/mcp.json — the same tool list this door serves.
- POST https://theagentsindex.com/api/submit (JSON: url, name, note?) — requires a signed-in account with verified email; send its session Cookie header, including on MCP submit_listing requests. Sign in or explicitly create an account at https://theagentsindex.com/login?next=%2Fsubmit, then verify its email. No session → HTTP 401 {error:'sign_in_required', message, loginUrl}; unverified session → HTTP 403 {error:'email_unverified', message, email, remediation}. Company gate: the account's address, or an address it proved with a code, must be on the submitted website's domain; otherwise HTTP 403 {error:'domain_unverified', message, domain, remediation} — POST /api/domain-proof/start {url, email} mails a 6-digit code to an address at that domain, POST /api/domain-proof/verify {proofId, code} redeems it, then submit again. Personal mailboxes (gmail, outlook…) never prove a domain. Refusals create no account, mail or listing. Identity/email come only from the session. Success returns 201 {id, slug, status:'Draft'}; research + the quality gate decide publication. Your own (url, session email) replay within 24h returns 200; validation returns 400 and duplicate/claim responses 409. Track progress at GET https://theagentsindex.com/api/submissions/{id}. The https://theagentsindex.com/submit form uses the same account-first flow.
- Claim a listing you own at https://theagentsindex.com/claim/{slug} — email-domain verification unlocks the owner dashboard (edits, tier upgrades).
- Category answer sheets at https://theagentsindex.com/category/{slug}.md (+ structured export at https://theagentsindex.com/category/{slug}.json).
- https://theagentsindex.com/.well-known/mcp.json + https://theagentsindex.com/.well-known/agents.json — machine discovery: the MCP door and the narrowed REST operations with I/O descriptors.
- https://theagentsindex.com/index.json — the machine manifest: every surface here, live counts, license, attribution.
- https://theagentsindex.com/license.xml — the RSL 1.0 licence document (application/rsl+xml): what machines may and may not do with this corpus, in machine-readable form.
- https://theagentsindex.com/openapi.json — OpenAPI 3.1 for the read + submit endpoints.
- POST https://theagentsindex.com/api/lead (JSON: name, email, message>=20, consent:true, listingSlug|categorySlug) — send a buyer-intent enquiry to a agent or category. Keyless, CSRF-free as application/json. Returns { ok, delivered }.
- POST https://theagentsindex.com/api/newsletter/subscribe (form-encoded: email; needs a same-origin Origin header) — single opt-in to the digest. Returns an HTML confirmation.
- https://theagentsindex.com/sitemap.xml — every indexable page.
- https://theagentsindex.com/api/search?q=… — full-text + faceted search over the Published corpus (JSON).
- Every listing page carries schema.org JSON-LD (SoftwareApplication, Organization, WebSite, BreadcrumbList) — parse it for citable structured facts, and /{slug}.md for the full sheet.
- Research/analysis twins at https://theagentsindex.com/blog/{slug}.md (index at https://theagentsindex.com/blog.md), sourced reports as answer-first markdown.
- Methodology (how agents are researched & quality-gated) at https://theagentsindex.com/methodology.md
- Guide twins at https://theagentsindex.com/guides/{slug}.md, each researched buyer's guide as answer-first markdown — the full body without the page chrome.
- Collection twins at https://theagentsindex.com/collections/{slug}.md, a curated set with its selection claim, how it was chosen, and a members facts table.
- Comparison twins at https://theagentsindex.com/compare/{a}-vs-{b}.md, the head-to-head facts table + verdict as markdown.
- Ranked best-of twins at https://theagentsindex.com/best/{category}/{tag}.md, one ranking per category-and-tag intersection as markdown.
- Per-listing alternatives twins at https://theagentsindex.com/{slug}/alternatives.md, the "X alternatives" comparison as markdown.
- The flagship "Best AI Agents" ranking as markdown: https://theagentsindex.com/best-ai-agents.md
- "The Best AI Support Agents (2026)" ranking as markdown: https://theagentsindex.com/best-ai-support-agents.md
- "The Best AI Voice Agents (2026)" ranking as markdown: https://theagentsindex.com/best-ai-voice-agents.md
- "The Best AI Research Agents (2026)" ranking as markdown: https://theagentsindex.com/best-ai-research-agents.md
- "The Best AI Sales & Marketing Agents (2026)" ranking as markdown: https://theagentsindex.com/best-ai-sales-marketing-agents.md
- "The Best AI Agent Platforms (2026)" ranking as markdown: https://theagentsindex.com/best-ai-agent-platforms.md
- "The Best AI Coding Agents (2026)" ranking as markdown: https://theagentsindex.com/best-ai-coding-agents.md
- "The Best AI Agent Frameworks (2026)" ranking as markdown: https://theagentsindex.com/best-ai-agent-frameworks.md

## Terms

Reading is free and keyless. Use is licensed, not unconditional: The Agents
Index Machine Access Terms 1.0 (https://theagentsindex.com/terms): permitted
machine use — search, ai-input; prohibited — ai-train. Attribution required:
Data from The Agents Index (https://theagentsindex.com). Attribution
required: name The Agents Index and link https://theagentsindex.com when
citing.

Machine-readable terms: https://theagentsindex.com/license.xml — RSL 1.0
(application/rsl+xml), also announced as a License: directive in
https://theagentsindex.com/robots.txt. Human-readable text:
https://theagentsindex.com/terms.

Agents may write as well as read: POST /api/submit and MCP submit_listing
require a verified account session Cookie header. Sign in or create an
account explicitly at https://theagentsindex.com/login?next=%2Fsubmit.
Missing session returns 401 sign_in_required with loginUrl; unverified email
returns 403 email_unverified with remediation; an account not on the
website's domain gets 403 domain_unverified and proves an address there via
/api/domain-proof/start and /verify. Refusals create no account, mail or
listing. Successful Drafts are inert until research + the quality gate clear
them. Checking a submission is open (GET /api/submissions/{id}), and a claim
starts at /claim/{slug}.