AI Agents & Skills
You can operate your entire print farm through an AI agent. Printago is built so an agent (Claude, ChatGPT, a custom tool-use loop, or similar) can list printers, queue jobs, manage orders, and troubleshoot the queue the same way a person would, with structured output it can parse reliably.
There are two complementary pieces: the CLI is designed to be agent-friendly and ships as a Claude Code skill, and these docs publish machine-readable artifacts an agent can fetch directly.
The CLI is built for agents
Every feature of the Printago CLI is designed to be driven by an agent as easily as by a person.
- Structured output. When the CLI's output is piped (not a terminal), every command returns a single-line JSON envelope:
{ ok, command, method, path, status, meta?, result }. Errors return{ ok: false, error, fix }with a suggested fix. An agent parses stdout directly, with no human text to scrape. - Discoverable. An agent never has to guess. Running
printagoprints the full command tree, and<command> --helpresolves the request-body schema into a field list, so the agent rarely needs to read the API spec to build a request. - Action hints.
printago hints <object>returns the API's own suggestions for what to do with an object (for example, the actions available on a printer or a print job), each mapped to the exact command that performs it.
# piped output is machine-readable
printago printers list | jq '.result[].name'
# discover what you can do with a printer, mapped to commands
printago hints printer
Claude Code skill
The CLI ships as a Claude Code skill. Once it is installed, point Claude at your project and ask it to manage your farm in plain language ("show me which printers are idle", "requeue the failed jobs from order 1234"). Claude runs the printago commands directly and reads the structured results. The skill teaches Claude the command conventions, the safe way to handle your credentials, and the common multi-step workflows (such as upload to part to build), so it does not rediscover them each session.
There are two ways to install it.
With the CLI. If you installed the CLI, it can install its own skill:
printago skill install
This writes the skill to ~/.claude/skills/printago/, where Claude Code picks it up in every project. Start a new Claude Code session (or run /reload-plugins) and you are ready.
As a plugin. You can also install it from the Printago plugin marketplace, which does not require the CLI on its own:
/plugin marketplace add printago/printago-skill
/plugin install printago-skill@printago-skill
Either way, authenticate the CLI first with printago auth login so the skill has access to your store.
Machine-readable docs and API
For agents that call the REST API directly or navigate the documentation rather than using the CLI, these docs ship machine-readable artifacts so an agent can understand the product without crawling page-by-page.
| File | URL | What it's for |
|---|---|---|
llms.txt | docs.printago.io/llms.txt | Single-fetch index of every doc page with a one-line description, following the llms.txt convention. Start here. |
docs_map.md | docs.printago.io/docs_map.md | Hierarchical map of every page with all its section headings. Fetch this when the agent needs to decide which pages to read. |
sitemap.xml | docs.printago.io/sitemap.xml | Standard sitemap listing every page URL. |
| OpenAPI spec | docs.printago.io/printago-api-swagger.json | OpenAPI 3.1 spec for the REST API. Feed this to an agent that needs to call Printago endpoints. |
| Developers site | developers.printago.io | Interactive API reference and guides. |
All of these are plain text or JSON, live at stable URLs, and are regenerated whenever the docs are rebuilt.
Pointing an AI agent at the docs
If you're using an AI coding agent to build on top of Printago, paste a few lines like this into your prompt so it knows where to look:
For Printago product documentation, fetch https://docs.printago.io/llms.txt first.
Follow up with https://docs.printago.io/docs_map.md or specific pages from
https://docs.printago.io/docs/* as needed.
For the Printago REST API, fetch the developer AI resources at
https://developers.printago.io/llms.txt, or the full OpenAPI spec at
https://docs.printago.io/printago-api-swagger.json.
Any agent with a web-fetch tool can use these files directly: give it the llms.txt URLs as a starting point and let it pull in whichever pages it needs.
If you reference the docs frequently, add the same lines to your project's agent instructions file instead so you don't have to repeat them. Most tools read an AGENTS.md file in your repo; Claude Code reads CLAUDE.md instead.
Feedback
If you're building something on top of Printago and want different discovery formats (e.g. llms-full.txt with the entire corpus inlined, per-section indexes, embeddings), let us know in Discord.