# Ablo developer resources

Every Ablo developer surface in one place: the @abloatai/ablo TypeScript SDK, the REST API reference and OpenAPI document, the hosted MCP server, the ablo CLI, and the llms.txt manifests agents read.

Everything you need to build on Ablo, at a fixed address. Humans can read every page here; agents can fetch every one of them as Markdown, JSON, or plain text.

## Build with the SDK

The typed client, and the guides that teach it.

- [@abloatai/ablo — TypeScript SDK](https://www.npmjs.com/package/@abloatai/ablo) — The typed client for agents, apps, and servers. Declare a schema, get ablo.<model> reads and writes, claims, and live deltas.
- [Ablo documentation](https://docs.abloatai.com) — Quickstart, guides, and the full reference for the SDK, CLI, and engine.
- [Ablo quickstart](https://docs.abloatai.com/quickstart) — Install, scaffold, push a schema, and make your first coordinated write.
- [Ablo CLI (ablo)](https://docs.abloatai.com/cli) — Scaffold a project, push a schema, connect a database, and mint branch-scoped keys.
- [Ablo source on GitHub](https://github.com/Abloatai) — The open SDK, its examples, and the changelog.

## Call the API directly

For runtimes with no SDK, and for generating your own client.

- [Ablo REST API reference](https://docs.abloatai.com/api) — Every endpoint, its parameters, and its typed error codes.
- [Ablo OpenAPI document](https://www.abloatai.com/openapi.json) (`application/json`) — The machine-readable OpenAPI 3.1 description of the same API, generated from the schemas the server validates against. Feed it to a code generator.

## Connect an AI assistant

Model Context Protocol, for an agent that works inside an assistant.

- [Ablo MCP server (ablo)](https://www.abloatai.com/mcp) (`application/json`) — Docs, schema validation, and scaffolding for the Ablo SDK.
- [Ablo MCP server manifest](https://www.abloatai.com/.well-known/mcp.json) (`application/json`) — The server.json registry manifest, so a client with only the domain can find the endpoint and its transport.
- [MCP integration guide](https://www.abloatai.com/mcp) — How to expose your own resources as typed MCP tools with coordination intact.

## Read the site as an agent

Every page on this domain answers to Accept: text/markdown, and these files describe the whole of it.

- [llms.txt](https://www.abloatai.com/llms.txt) (`text/plain`) — The short manifest of what Ablo is and where its documentation lives.
- [llms-full.txt](https://www.abloatai.com/llms-full.txt) (`text/plain`) — The entire published documentation corpus in one fetch.
- [Documentation as plain Markdown](https://www.abloatai.com/api/docs/quickstart) (`text/markdown`) — Any documentation page by slug, served as text/markdown — for example /api/docs/quickstart.
- [sitemap.xml](https://www.abloatai.com/sitemap.xml) (`application/xml`) — Every indexable page on this domain.

## Fetch any page as Markdown

Every page on this domain serves a Markdown representation from its own URL. Ask for it with an Accept header, or add a .md suffix if your client cannot set one. Responses carry Vary: Accept, so a cache keeps the two representations apart.

```bash
curl -H "Accept: text/markdown" https://www.abloatai.com/developers
```

---

Canonical URL: https://www.abloatai.com/developers

Also machine-readable on this domain: /llms.txt, /llms-full.txt, /openapi.json, /sitemap.xml, /robots.txt