AI & MCP
Use Tomba docs with AI assistants
The Tomba documentation is also published as Markdown files that AI assistants and coding tools can read. Give an assistant the files that cover your task when you ask it to write code against the Tomba API.
To let an assistant call Tomba instead of reading about it, connect the remote MCP server.
Files
| File | Contents |
|---|---|
llms.txt | An index of the documentation: one line per page with its title, its description, and a link to its .md file. It doesn't contain the pages themselves. |
llms-full.txt | Every documentation page in one Markdown file. It doesn't include the API reference. |
Page URL + .md | One page as Markdown. Append .md to the URL of any documentation page, for example https://docs.tomba.io/authentication.md. |
openapi.md | The API reference as Markdown: every endpoint with its parameters, request bodies, and response schemas. |
Give the files to an assistant
- Choose the files the task needs: the
.mdpages for the features you use, andopenapi.mdwhen the assistant needs exact parameters and response fields. Usellms.txtto find the right pages. - If the assistant can fetch URLs, paste the file URLs into your request. Otherwise, download the files and attach them, or add them to the files your tool keeps as project context.
- Ask for the code. Name the endpoints you want and the language, and ask the assistant to send the key and secret in the
X-Tomba-KeyandX-Tomba-Secretheaders, as described in Authentication.
llms-full.txt and openapi.md are large files, and many assistants can't load either one whole into a conversation. When that happens, give the assistant individual .md pages instead.
Last modified on