# Taskade — AI workspace that turns one prompt into a live app with agents and automations wired in. > Source: The Agents Index — https://theagentsindex.com/taskade (structured, researched, re-verified) Taskade is an AI-native workspace where a plain-English prompt produces a hosted app with a database, client sign-in, AI agents and automations. Its Workspace DNA layers projects, agents and automation runs, reachable from REST, Action API and MCP. **Best for:** Operators and agencies shipping client CRMs, portals and dashboards without developers | Fact | Value | | --- | --- | | Website | https://taskade.com | | Category | Agent builders, Agent platforms | | Open source | No | | Price | Free | | Price source | https://www.taskade.com/pricing | | Price checked | 26 October 2026 | | MCP server | Partly | | API | Not stated | | Self-host | Not stated | | Facts | 21 facts from 6 sources | | Facts checked | 26 October 2026 | Screenshot of Taskade: https://theagentsindex.com/img/1920/screenshots/taskade.webp, captured 27 October 2026 **What we checked** - Pages last read: 26 October 2026 - Facts checked: 21 facts (From 6 sources) - AI answers: Named in 1 of 151 (Our October check) ## What we checked on Taskade ### How much it does unattended **Multi-agent** · Partly — AI Teams send one task to several agents with Auto, Everyone or Orchestrate execution modes, and agents can chain through automations, but the agent-team step requires Pro or above. > AI Teams (**Ask an agent team**) | **Pro and above** > — https://www.taskade.com/docs/guides/autonomous-agents · Plan Gating table, AI Teams row, checked 26 October 2026 **Agent permissions** · Partly — Agents run from automation triggers rather than per-step prompting, but the agent-team step refuses tools that need manual approval and an agent cannot reach an arbitrary external MCP server without an automation in between. > The step does not support Agent App tools. If a tool needs manual approval, the run fails. > — https://www.taskade.com/docs/guides/autonomous-agents · Pattern: Multi-Agent Team, Execution modes section, checked 26 October 2026 ### Which models it runs on **Open models** · Yes — The TSK-1 layer routes to open-weight models alongside closed ones, with DeepSeek and Qwen named, and a separate open-source-value model class for cheap routine steps. > TSK-1 uses frontier models from top AI labs, including OpenAI GPT and open-weight models such as DeepSeek and Qwen. > — https://www.taskade.com/pricing · FAQ, "Which AI models does Taskade use?", checked 26 October 2026 **Model choice** · Yes — TSK-Auto routes each turn by default, but you can set a TSK rung per surface or pin one exact model per agent with Custom Engine, so the choice is the buyer's when they want it. > Custom Engine | One exact model | Comparing models, or one model suits the task > — https://www.taskade.com/docs/models · Overview table, Custom Engine row, checked 26 October 2026 ### Building with it **SDK languages** · Partly — Python and cURL examples plus a typed TypeScript fetch wrapper are documented, but the TypeScript client @taskade/sdk is a preview that is not on npm and covers only part of the v2 surface. > `@taskade/sdk` is not published on the public npm registry. `npm install @taskade/sdk` returns `404 Not Found` today. Use `fetch`, as the examples on this page do. > — https://www.taskade.com/docs/guides/typescript-sdk · Important callout above "Before you start", checked 26 October 2026 **Tracing and evals** · Yes — Each automation has a Runs tab with filterable run history, credit and token usage breakdowns per account and per app, and automatic shutdown after five consecutive failures. > The automation's **Runs** tab shows the **Run History**. Filter it with **All**, **Completed**, **Failed**, or **Running**. > — https://www.taskade.com/docs/guides/autonomous-agents · Monitoring and Observability section, Run history, checked 26 October 2026 ### What it costs to run **How it meters** · Yes — Billing is a flat monthly workspace fee that includes a monthly AI credit allowance, with credits consumed per AI build, agent run and AI automation step and extra packs bought as needed. > Credits power all AI work in Taskade: generating apps, running agents, and executing AI steps in automations. > — https://www.taskade.com/pricing · FAQ, "How do AI credits work?", checked 26 October 2026 **Cost of one job** · Partly — One job is one app build, which the vendor puts at a few thousand credits against the 50,000 monthly credits of the cheapest paid plan, but it publishes no credit-to-dollar rate so the cost per build is not exact. > A typical Taskade Genesis app build uses a few thousand credits, so each paid plan's monthly allowance covers multiple builds and refinements. > — https://www.taskade.com/pricing · FAQ, "How far do my monthly credits go?", checked 26 October 2026 **Free tier** · Yes — A standing $0 Free plan is self-serve from the signup page with no sales call, capped at three live apps, one agent, two people and ten automation runs a month. > Free For your first apps $0/mo > — https://www.taskade.com/pricing · Plan cards, Free column price cell, checked 26 October 2026 **MCP server** · Partly — A hosted MCP server at taskade.com/mcp lets Claude, Cursor, VS Code or ChatGPT operate a workspace over Streamable HTTP with OAuth 2.0 and PKCE, but inbound MCP needs any paid plan and is absent on Free. > Hosted Taskade MCP | Every paid plan (Starter and above) | The `mcp.access` plan feature gates each workspace > — https://www.taskade.com/docs/mcp · Access table, Hosted Taskade MCP row, checked 26 October 2026 **Bring your own key** · Partly — Your own OpenAI, Anthropic, Google Vertex or Gemini key runs custom agents and automation AI steps from Pro up, with xAI, Groq, Mistral and more from Business, but EVE builds still bill Taskade credits. > Yes. Custom AI Agents run on your own OpenAI, Anthropic, or Google Vertex AI key on Pro and above, and on your own xAI, Groq, Mistral, or Together AI key on Business and above. > — https://www.taskade.com/pricing · FAQ, "Can I use my own AI key?", checked 26 October 2026 ### Buying it for a team **A company can buy it** · Yes — Business, Max and Enterprise are checkout links an organisation can buy online, with SAML SSO, unlimited people, SCIM on Enterprise and volume terms available from sales. > Unlimited people and SAML single sign-on > — https://www.taskade.com/pricing · Plan cards, Business column feature list, checked 26 October 2026 **Seat model** · Partly — Plans are flat monthly fees rather than per-seat prices, but each carries a people ceiling: two on Free, ten on Pro, unlimited from Business up, with app visitors not counted. > 10 people and 100+ integrations > — https://www.taskade.com/pricing · Plan cards, Pro column feature list, checked 26 October 2026 **Pooled budget** · Yes — Credits and automation runs are workspace-level balances rather than per-seat allowances, so a Business team draws its 150,000 monthly credits and 10,000 runs from one shared pool. > Shared credit pool (150,000/month) > — https://www.taskade.com/pricing · FAQ, "Which plan is best for teams building internal tools?", checked 26 October 2026 **Admin controls** · Yes — Role-based access with Owner-through-Viewer roles and workspace scoping ships on every plan, and generated apps inherit the parent workspace's access rules rather than needing their own setup. > Role-based access | All plans | Granular roles, Owner through Viewer > — https://www.taskade.com/security · Enterprise Access Controls table, Role-based access row, checked 26 October 2026 **Audit log** · Partly — Taskade logs account and workspace activity and will export it on request, but says plainly it offers no immutable logs, no tamper detection and no prebuilt compliance report formats. > we log account and workspace activity, monitor security events, and can export workspace activity for you. We do not currently offer cryptographically immutable logs, tamper detection, or prebuilt compliance report formats. > — https://www.taskade.com/security · Compliance and auditing section, Audit trail paragraph, checked 26 October 2026 **Single sign-on** · Partly — SAML 2.0 with OIDC support covers Okta, Azure AD and Google Workspace from the Business plan up, and SCIM 2.0 provisioning stays on Enterprise, so Free and Pro have neither. > SSO / SAML 2.0 | Business and above | Okta, Azure AD, Google Workspace, and other major identity providers; OIDC supported. > — https://www.taskade.com/security · Enterprise Access Controls table, SSO / SAML 2.0 row, checked 26 October 2026 ### What happens to your code **Opt out of training** · Yes — No opt-out is needed because the default is no training: providers are contractually barred from training on customer content, though the vendor disclaims liability for third-party compliance. > **Zero Training Policy**: We contractually prohibit AI providers from using your data to train general AI models. > — https://www.taskade.com/security · AI & Taskade Genesis Security section, Zero Training Policy, checked 26 October 2026 **Data residency** · Partly — Workspace data sits in one AWS region in the United States with no published region choice, and AI requests may be routed to model providers operating outside the US, including in China. > Deployment | Single AWS region (US East) > — https://www.taskade.com/security · Business continuity table, Deployment row, checked 26 October 2026 **Getting out** · Yes — Pro and above export an app to a GitHub repository, any plan can download an app as a .tsk bundle, and purchased or earned Extra Credits are stated not to expire. > Yes. On Pro and above, open the Publish menu and choose Export to GitHub. Taskade creates a GitHub repository with your app and its projects, agents, and automations. > — https://www.taskade.com/pricing · FAQ, "Can I export my Taskade Genesis app code?", checked 26 October 2026 **Certifications** · Partly — Only Google CASA is held today; the vendor states SOC 2 Type II has no audit report yet and that GDPR, CCPA and ISO 27001 work is in progress, with no signed DPA or SCCs offered. > SOC 2 Type II | In progress; current practices are designed to align with SOC 2 requirements but no audit report exists yet > — https://www.taskade.com/security · Compliance & Certifications table, SOC 2 Type II row, checked 26 October 2026 ## Taskade price and terms **Terms** - Free plan: Yes — Free is $0/mo: up to 3 live apps with hosting and a database, 1 AI agent, 10 automation runs a month, 2 people, 250 MB storage, 7-day version history, and 0 monthly AI credits (earn through referrals and social actions, or buy a pack). **What needs a higher plan** - Pro: Multi-agent, Bring your own key - Business: Single sign-on ## For AI agents - MCP server: Partly Pricing page as we captured it: https://theagentsindex.com/img/1920/screenshots/taskade/pricing.webp, captured 27 October 2026 Read on https://www.taskade.com/pricing, 26 October 2026. ## Taskade alternatives - [Blits.ai Agentic AI for the Enterprise](https://theagentsindex.com/blits-ai-agentic-ai-for-the-enterprise) · Price on request: Enterprise platform for governed agentic workflows across models, tools, data and systems. - [ElevenLabs Agents](https://theagentsindex.com/elevenlabs-conversational-ai) · Free plan, then $6 a month: Hosted platform for building voice and chat agents, with telephony, SDKs and a choice of LLM. - [Dust](https://theagentsindex.com/dust) · Free plan, then €30 a seat a month (about $34): Platform for building and running company agents on connected tools, data and 20+ models. - [Gumloop](https://theagentsindex.com/gumloop) · From $37 a month: Multiplayer AI agent builder: teams build agents over their apps while IT keeps governance and spend control. - [Dify](https://theagentsindex.com/dify) · Free plan, then $59 a month: Visual platform for agentic workflows and RAG, on Dify Cloud, your VPC or self-hosted. - [Manus](https://theagentsindex.com/manus) · Free plan, then $20 a month: General-purpose AI agent that plans and executes multi-step tasks on its own cloud computer. - [Letta](https://theagentsindex.com/letta) · Free plan, then $20 a month: Stateful agent harness, cloud and SDK for agents with persistent, git-versioned memory. - [Genspark](https://theagentsindex.com/genspark) · Free plan, then $24.99 a month: All-in-one AI workspace with an autonomous Super Agent plus agents for slides, docs, media and code. - [PitchAI Agent Engine](https://theagentsindex.com/agent-engine) · From €400 a month (about $447): Scheduled agents that run recurring office administration across accounting, CRM and mail systems. 8 of the 9 publish a price. [All 9 alternatives](https://theagentsindex.com/taskade/alternatives/) ## Questions about Taskade ### How much does Taskade cost? Taskade is free to use. From taskade.com/pricing, checked 26 October 2026. ### Does Taskade have an API or an MCP server? It has a limited MCP server. ### What have we checked on Taskade? 21 facts, under these headings: How much it does unattended, which models it runs on, building with it, what it costs to run, buying it for a team, what happens to your code. ### How do we check Taskade? We hold 21 facts on Taskade, quoted from 6 pages. Facts and price last checked 26 October 2026. How we check every listing: https://theagentsindex.com/methodology/ ## Sources and changes Last changed 27 October 2026 **Sources on taskade.com** (6) - https://www.taskade.com/pricing, read 26 October 2026 - https://www.taskade.com/security, read 26 October 2026 - https://www.taskade.com/docs/guides/autonomous-agents, read 26 October 2026 - https://www.taskade.com/docs/guides/typescript-sdk, read 26 October 2026 - https://www.taskade.com/docs/mcp, read 26 October 2026 - https://www.taskade.com/docs/models, read 26 October 2026 ## Is this yours? AI assistants like ChatGPT and Claude opened The Agents Index pages 2,746 times in September. Claim Taskade: https://theagentsindex.com/claim/taskade Report an error if a fact is wrong: hello@theagentsindex.com, subject "Report an error: Taskade (taskade)".