Mintlify
Also known as: mintlify.com, Mintlify agent, Mintlify Autopilot, Mintlify Index, Mintlify MCP server
Mintlify is a documentation platform for people and AI agents, hosting docs kept in Git for more than 20,000 companies with llms.txt, a hosted MCP server per site and an AI assistant. Its agents keep docs current: the Mintlify agent writes changes from Slack, Teams or a prompt, and Autopilot watches merged code, support questions and Slack threads, drafting updates as pull requests and learning the team's style. Pro costs $450 a month billed yearly with 10,000 credits.
Mintlify builds a documentation platform for people and AI agents alike. It hosts docs kept as Markdown or MDX in Git for more than 20,000 companies, including Microsoft, Anthropic, Amazon, Coinbase, Cognition, AT&T and Rivian along with startups such as Lovable, Replit, Perplexity and Harvey, and served more than 300 million visitors and 2 billion agents in the past year.
Anthropic's docs draw more than 10 million visits a month, Coinbase deploys its docs 50 times faster, and HubSpot builds three times faster with half the engineering resources. Every site serves llms.txt and Markdown to agents and gets a hosted MCP server, and Mintlify Index pools the docs of more than 5,000 Mintlify sites into a retrieval layer for coding agents.
Three agents keep the docs current. The Mintlify agent works from Slack, Microsoft Teams, the dashboard or a custom app: given a prompt, a pull request, a Slack thread or a file, it reads source code in any connected repository and context from Slack, Notion, Linear and Jira, plans and writes the change across files, runs build checks and opens a pull request.
Autopilot, in beta on paid plans, watches for change on its own, reading merged code diffs, repeated support questions and Slack threads that flag a mismatch, drafts the updates and tags reviewers, and it learns the team's tone and release habits from every pull request merged or closed. On Enterprise, automations defined in the repository or dashboard run set jobs such as updating pages from code changes, drafting changelogs and translating content. Nothing ships without review: every change arrives as a pull request and no agent commits to the main branch. For readers, an AI assistant answers questions from the docs and connects to support tools.
The company keeps a SOC 2 Type II report and an ISO 27001 certificate in its trust center. Enterprise adds single sign on, SCIM, role based permissions, audit logs and IP allowlists, along with EU hosting and self hosting in the customer's own cloud or on premises on AWS or Kubernetes. The Free plan covers five editors with hosting, search, a custom domain and the MCP server. Pro costs $450 a month billed yearly or $540 monthly with unlimited editors and 10,000 credits a month, where an assistant answer uses 25 credits and an Autopilot pull request 250, and Enterprise is quoted with custom credit volume. New accounts get Pro free for 14 days.
The natural customers are software companies whose product moves faster than their docs, especially those whose readers are increasingly coding agents rather than people. A team cannot pick the model behind the agents, and drafts are judged in human review of each pull request rather than against a test set before they arrive.
Vendor details
Canonical URL
https://www.mintlify.com
Category
Agent infrastructure
Subcategory
Documentation platform for people and AI agents, with an agent and an always on Autopilot that keep docs current through pull requests, plus MCP servers, llms.txt and a docs retrieval index for coding agents
Funding status
Venture backed private company.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
GitHub (Mintlify GitHub App) and the customer's own Git provider; Slack and Microsoft Teams for the agent; Notion, Linear and Jira as agent context; support ticket tools and feature flags for Autopilot on Enterprise; support integrations for the assistant; hosted MCP servers, WebMCP and llms.txt; developer API, admin APIs, webhooks and websockets; analytics data warehouse streaming.
In practice
An API company merges a pull request that adds a new pagination parameter. Autopilot reads the diff, drafts updates to the endpoint reference and the guide that uses it, tags the developer relations lead, and the change goes live once she merges it.
A support engineer notices the same setup question five times in a week and tags the Mintlify agent in the Slack thread. The agent checks the source code, writes a troubleshooting page and opens a pull request linked to the thread.
A developer tools company with docs in English turns on the translation automation, so each merged change to the English pages produces matching pull requests for its Japanese and German sites.
Sources & related URLs
Related / legacy domains
Agentic Index coverage score
10.5 / 14 capabilities · 75%
| Integrations & Tool Calling | Full |
|---|---|
|
The agent reads source code in any repository with the Mintlify GitHub App installed and pulls live context from Slack, Notion, Linear and Jira, then opens pull requests in the docs repository. Autopilot also draws on support tickets and feature flags on Enterprise, the assistant connects to support tools, and the agent can be reached from Slack, Microsoft Teams or a custom app. Sourcemintlify.com/docs/agentread 2026-11-19 |
|
| Workflow Orchestration | Full |
|
Given a task, the agent researches the codebase and existing pages, plans the edit, writes changes across several files, runs build checks and opens a pull request. Automations turn this into standing jobs: updating pages from code changes groups related changes into one pull request and folds later changes into any unmerged one, while others draft changelogs or translate content. Sourcemintlify.com/docs/automations/reference.mdread 2026-11-19 |
|
| Knowledge Grounding & RAG | Full |
|
Drafts are grounded in the team's existing pages, the source code of connected repositories and context from Slack, Notion, Linear and Jira, and Enterprise adds external sources for the assistant. Readers get assistant answers drawn from the docs themselves, and Mintlify Index pools the docs of more than 5,000 Mintlify sites into one retrieval layer that coding agents can query. Sourcemintlify.com/docs/agentread 2026-11-19 |
|
| Human Oversight & Guardrails | Full |
|
Every change the agents make arrives as a pull request, and none commits to the main branch. Autopilot tags team members and emails them to review, nothing ships until someone merges, teams set the agent's review process in Slack or Teams settings, and a hard spending limit or an overage switch caps how many credits the agents can use in a month of work. Sourcemintlify.com/docs/agentread 2026-11-19 |
|
| Security, Identity & Governance | Full |
|
Mintlify keeps a SOC 2 Type II report and an ISO 27001 certificate in its trust center, along with a data processing agreement, architecture diagram and system description. Enterprise adds single sign on for docs and dashboard, SCIM provisioning, role based permissions, audit logs and session IP allowlists, and any docs site can sit behind reader authentication. Sourcesecurity.mintlify.comread 2026-11-19 |
|
| Observability & Auditability | Partial |
|
Platform and agent analytics on Pro show how readers and AI agents use the docs, with user feedback, and Enterprise streams analytics to a data warehouse and keeps audit logs. The Autopilot page lists the suggestions and pull requests it has made, but there is no trace of each agent run, its sources or its cost beyond the pull request itself and the credit count. Sourcemintlify.com/pricingread 2026-11-19 |
|
| Memory & State Persistence | Full |
|
Autopilot keeps memory of how a team works: its tone and voice, and whether docs change with each release or right away. It tunes itself from every pull request a team merges or closes, and a team can steer it in conversation, so later drafts follow lessons from earlier reviews without anyone maintaining a style prompt. Automations, by contrast, run on the prompt their owner keeps. Sourcemintlify.com/blog/autopilotread 2026-11-19 |
|
| Deployment & Data Residency | Full |
|
Enterprise customers can self host in their own cloud or on premises on AWS or a Kubernetes platform, including Azure and Google Cloud, or choose EU hosting for their docs. Free and Pro sites run on the hosted platform at 99.99% uptime, with the docs themselves stored as Markdown or MDX in the customer's own Git repository rather than locked in the platform. Sourcemintlify.comread 2026-11-19 |
|
| Prebuilt Agents / Templates / Packs | Full |
|
Autopilot comes ready on Pro and needs no setup beyond connecting sources, and the writing agent and reader assistant arrive switched on rather than assembled from parts. Enterprise adds predefined automations for updating pages from code changes, drafting changelogs and translating content, each with defaults a team can change, and Pro adds agent skills, AI translations and support integrations. Sourcemintlify.com/docs/automations/reference.mdread 2026-11-19 |
|
| Triggers & Channel Coverage | Full |
|
Autopilot starts work when something changes, reading merged code diffs, spotting repeated support questions and picking up Slack threads where someone flags that the docs are wrong, with no prompt needed. Automations run on code changes and other events defined in the repository, and people start the agent from Slack, Teams, the dashboard or an app it is embedded in, with reviewers alerted by tag and email. Sourcemintlify.com/blog/autopilotread 2026-11-19 |
|
| Model Flexibility & Routing | Not documented |
|
Mintlify alone picks and runs the models behind the agent, Autopilot and the assistant. There is no setting to choose a model, bring a provider key or route a task to a cheaper or stronger model; the controls on offer cover credits, spending limits and what each agent may do rather than which model does the writing or answering. Sourcemintlify.com/pricingread 2026-11-19 |
|
| APIs / SDKs / MCP Extensibility | Full |
|
A hosted Mintlify MCP server gives coding assistants access to the dashboard and content, and every docs site gets its own server for agents plus WebMCP, llms.txt and Markdown. Pro adds a developer API, admin APIs, webhooks and websockets, the agent can be embedded in custom apps, and a command line tool previews and checks docs on a developer's machine. Sourcemintlify.com/pricingread 2026-11-19 |
|
| Testing, Debugging & Optimization | Not documented |
|
The agent runs build checks before opening a pull request, and Pro adds grammar and spelling checks and preview deployments, which test the docs rather than the agents. There is no way to score drafts against a reference, replay an agent on past code changes or compare settings side by side, so quality is judged in human review of each pull request as it arrives. Sourcemintlify.com/docs/agentread 2026-11-19 |
|
| Browser / Computer-use | Not documented |
|
The agents work through Git, connectors and messaging apps, reading code and threads through integrations and writing through pull requests. None of them drives a browser or a desktop to check a live product screen, capture a flow for a guide or follow steps in an app that has no integration, so that part of documentation work still falls to the writers. Sourcemintlify.com/docs/agentread 2026-11-19 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Free (5 editors); Pro $450/mo billed yearly or $540 monthly with 10,000 credits; Enterprise custom
Flat plan price with unlimited editors, plus credits consumed by the assistant, Autopilot and automations
Included quota
Pro: 10,000 credits a month; assistant answer 25 credits; Autopilot or automation update 250 credits; writing agent unlimited
What is public
Mintlify's Free plan covers five editors with hosting, search, a custom domain and the MCP server. Pro costs $450 a month billed yearly or $540 billed monthly, with unlimited editors and 10,000 credits a month, and Enterprise is quoted.
Billing mechanics
An assistant answer uses 25 credits and an Autopilot or automation pull request 250, runs that find nothing to change cost nothing, the writing agent is unlimited, and extra credits cost $0.01 each.
Cost watchouts
Autopilot pull requests at 250 credits each can use up the 10,000 monthly credits in 40 updates.
Variable cost rationale
The plan price is flat, but assistant answers and Autopilot pull requests draw credits, and usage past 10,000 credits bills at $0.01 each unless the team caps it.
Additional watchouts
Automations and self hosting are Enterprise only.
Overage / add-ons
$0.01 per credit on Pro, with discounted credit packs; hard spending limit or overage off available
Sales call required
Mixed (some tiers require a call)
Free / trial
Free plan for up to 5 editors without AI features; 14 day Pro trial with 5,000 credits
Lowest paid plan
Pro, $540 a month billed monthly or $450 a month billed yearly
Commercial notes
New accounts get Pro free for 14 days with 5,000 credits, open source projects can get Pro free, and startups can get half off Pro for a year.
Support SLA / resale
Enterprise adds a performance SLA, custom SLAs, Slack support, dedicated customer success and 24/7 incident monitoring.
Missing data
The Enterprise price and credit pack rates are not public.
Related vendors
- AgentOps — Agent observability and debugging platform: open source SDKs trace…
- AgentPhone — AgentPhone gives the AI agents a team builds their own US or…
- Agno — Python agent framework and AgentOS runtime (formerly Phidata) for…
- Airbyte — Airbyte moves data from operational systems into warehouses and…
- AIsa — Resource and payment gateway for AI agents, with one key to 110+…
- Albato — Albato is a no code integration platform that sells an embedded,…
Alternatives to Mintlify
Other products that do the work Mintlify does, and what each one is.