A developer portal your engineers and their AI assistants can both use
OpenDocs is developer portal software for guides, reference, and changelogs — published on your own domain, kept in two-way sync with GitHub, and queryable live by AI coding assistants through an MCP server.
What developers expect from a portal
A developer portal is the front door to your product for the people who build on it. When a developer lands there, they are not browsing — they have a job to do, and the portal either helps them finish it or sends them to a competitor's docs. That sets a high bar, and it is a different bar than a general knowledge base or an internal wiki has to clear.
The first thing developers expect is the right mix of content. Guides and tutorials get them from zero to a first working call. Reference documents every endpoint, parameter, and SDK method precisely, because reference is what they return to a hundred times after onboarding. A changelog tells them what moved and what broke between versions. A portal that has guides but no reference, or reference but no getting-started path, leaves half the audience stuck.
Second, developers expect the docs to live next to the code. Reference that is written by hand in a separate tool drifts out of date the moment an engineer ships a change. The docs that stay accurate are the ones engineers can edit in the same pull request as the code, in Markdown, under version control — while writers still get a friendly editor to polish the guides around them.
Third, developers expect fast search. Nobody reads a developer portal front to back. They arrive from a search engine or an internal link, hit the portal's own search, and expect to be on the right page in seconds. Slow or missing search is the fastest way to lose them.
And increasingly, developers expect their AI coding assistants to query the docs too. A growing share of developer work happens with an assistant open in the editor. If that assistant can read your live documentation instead of guessing from a training snapshot, its answers match how your product actually works today. A modern developer portal has to serve humans and their agents from the same source.
How OpenDocs builds that portal
Every capability maps to something developers expect — docs next to code, a branded hub, live search, and now docs their AI assistants can read.
GitHub two-way sync
Connect a space to a GitHub repository with a Personal Access Token and edits flow both ways. A push updates your pages through a webhook; a page save commits Markdown plus YAML frontmatter back to the repo. Engineers keep reference in the repo next to the code, writers polish the guides in the visual editor, and conflict detection shows a side-by-side comparison when both sides touch the same page. It is a real docs-as-code workflow without forcing your writers into git.
A branded portal on your domain
Publish the whole hub at your own custom domain with your colors and theme, structured navigation, and no generic subdomain. Built-in SEO — meta tags, sitemap, and canonical URLs — means your guides and reference are discoverable by the search engines developers actually start from. The portal looks like part of your product because it is.
Reader search built in
Every published portal ships with reader search, so a developer who lands mid-site can find the exact page they need without leaving it. Page feedback lets readers tell you when a doc missed the mark, so the gaps in your portal surface from the people using it.
API access and full API docs on Enterprise
The Enterprise plan adds API access and full API documentation, alongside analytics and insights, PDF and Markdown export, SSO/SAML, audit logs, and priority support — the pieces a larger developer-facing team needs to run a portal at scale and see how it is used.
The flagship: docs your AI coding assistants can query live
This is where an OpenDocs developer portal does something most portals cannot. Every space you publish is reachable through the OpenDocs MCP server — an endpoint built on the Model Context Protocol, the open standard for connecting AI assistants to real tools and data. Secure it with an OpenDocs API key, point a Claude Code or Cursor-class assistant at it, and your developer docs become a live, structured data source the assistant can read while a developer builds.
The server exposes four tools. list_spaces tells an agent which documentation sets it can reach. get_page_tree returns the full navigation structure of a space so the agent knows where things live. get_page returns the complete content of a page as clean plain text. search_pages runs a keyword search across a space. Together they let an assistant discover, navigate, read, and search your portal the same way a developer would — but at machine speed, in the middle of a coding session.
Because it reads your live published content over the streamable HTTP transport, there is nothing to export and no scraper to babysit. The instant you publish a change, the next query reflects it, so an assistant answers from your current API guides rather than a snapshot that predates your last release. It works with any MCP-compatible client — Claude Desktop, Claude Code, and a growing ecosystem of other MCP clients — and access is governed entirely by the API key you issue, which you can rotate to stop serving an agent at any time. Your published documentation doubles as a queryable knowledge source for AI agents, with no separate pipeline to maintain.
Being honest: there is no interactive try-it API console
We would rather set the right expectation than oversell. OpenDocs does not include an interactive try-it API explorer — the kind of runnable OpenAPI console that fires live requests from inside the docs and shows you the response. If that runnable playground is the core of what your developer portal needs, tools built around OpenAPI serve it better than we do, and we will point you to them: ReadMe is API-first with an interactive explorer, and Mintlify is docs-as-code with strong API reference tooling.
What OpenDocs is built for is the portal around your API: the guides that teach it, the tutorials that show it in context, the SDK docs, the concepts, the changelog, and the reference pages, all on a branded, searchable site that engineers and writers maintain together — and that AI assistants can query through MCP. For most teams that is the larger, harder half of a developer portal, and it is the half OpenDocs does exceptionally well. Many teams pair a dedicated API console with OpenDocs for everything else; a link from your portal to a hosted explorer is a small bridge to build.
Developer docs in Spanish and Portuguese
Reach Latin American and global developer audiences without running a separate localization project.
Up to 38 languages
AI Translations render your entire portal — guides, reference, and changelog — in up to 38 languages, including Spanish and Portuguese for the fast-growing developer communities across Latin America and Brazil.
Stays in sync with the source
Translations track the original. When you update an English page, the translated versions stay in step, so a Spanish or Portuguese developer never reads a stale copy of your docs while the English version has already moved on.
Bring your own API key
Translations run on your own Anthropic API key (BYOK), so AI usage is billed directly to your Anthropic account rather than marked up by OpenDocs. It is available on every plan, including the free trial.
Pricing that does not tax your readers
A developer portal is read by far more people than edit it, and OpenDocs pricing is built around that. Tiers are flat, with members included, and the developers and customers who read your published portal are never billed as seats — reader access is controlled by the space (public, or restricted for private docs), not by a per-seat license.
Pro is $55/month (or $45.65/month billed annually, $547.80/year) and includes 5 members, with additional members at $5/member/month monthly or $4/member/month annually. Enterprise is $99/month (or $82.50/month billed annually, $990/year) and includes 10 members, adding API access, full API docs, analytics and insights, PDF and Markdown export, SSO/SAML, audit logs, and priority support. Every plan — including the free trial — includes GitHub Sync, the MCP server, and AI features. The 14-day free trial needs no credit card and unlocks all Pro features so you can build a real portal before you decide. See the full pricing page for the complete breakdown.
Frequently asked questions
What is a developer portal?
A developer portal is the single place your developers go to understand and use your product: onboarding guides, tutorials, API and SDK reference, concepts, and a changelog, all under one branded, searchable site. OpenDocs is developer portal software that publishes that hub on your own domain, keeps it in two-way sync with a GitHub repository, and exposes it to AI coding assistants through an MCP server.
Does OpenDocs have an interactive try-it API console?
No. OpenDocs does not include an interactive try-it API explorer that sends live requests from the docs. It is the portal around your guides, tutorials, SDK docs, and concepts, with full API reference pages on Enterprise. If a runnable OpenAPI console is the core of what you need, tools like ReadMe or Mintlify are built around that, and we say so honestly.
How do engineers and writers work on the same docs?
GitHub Sync keeps a space and a repository in step in both directions. Engineers edit Markdown in the repo next to the code; writers use the block-based visual editor with no git knowledge required. A push updates the pages through a webhook, a page save commits Markdown plus YAML frontmatter back to the repo, and conflict detection shows a side-by-side comparison when both sides change the same page.
Can AI coding assistants read my developer docs?
Yes. Every published space is reachable through the OpenDocs MCP server, secured by an API key. MCP-compatible clients such as Claude Desktop and Claude Code can call list_spaces, get_page_tree, get_page, and search_pages to list, read, and search your live developer docs at query time, so an assistant answers from your current documentation rather than a stale snapshot.
Can I publish developer docs in more than one language?
Yes. AI Translations render your documentation in up to 38 languages, including Spanish and Portuguese for Latin American developer audiences. Translations run on your own Anthropic API key (BYOK) and stay in sync as you update the source, so a Spanish or Portuguese reader sees the same current content as the English original.
How much does a developer portal on OpenDocs cost?
OpenDocs uses flat tiers with members included. Pro is $55/month (or $45.65/month billed annually) with 5 members, and Enterprise is $99/month (or $82.50/month billed annually) with 10 members and adds API access, full API docs, analytics, PDF and Markdown export, SSO/SAML, and audit logs. A 14-day free trial needs no credit card, and readers of your published portal are never billed as seats.