LLMs.txt
LLMs.txt
How to get AI tools like Cursor, Windsurf, GitHub Copilot, ChatGPT, and Claude to understand APIpie’s AI routing platform, providers, models, and APIs.
What is LLMs.txt?
LLMs.txt is a structured documentation format designed specifically for Large Language Models (LLMs).
APIpie provides LLMs.txt files that expose canonical, machine-readable context about the APIpie platform so AI tools can reason correctly about:
- What APIpie does
- How AI routing works
- Which providers and models are supported
- How developers interact with the APIpie API
These files are optimized for AI consumption, not humans. They contain structured summaries, references, and links to authoritative documentation.
Available Routes
APIpie publishes the following LLM context files:
/llms.txtA concise, high-signal overview of the APIpie platform (~5–10K tokens). Intended for discovery, relevance checks, and standard context windows./llms-full.txtA comprehensive context file including routing concepts, provider behavior, API references, and deeper platform details (large context, 100K+ tokens).
Both routes are public, cacheable, and stable.
Choosing the Right File
/llms.txt.
It contains everything an AI tool needs to understand APIpie at a platform level and fits within standard LLM context limits.Use /llms-full.txt only if:- You need detailed API or routing behavior
- Your AI tool supports large context windows (100K–200K+ tokens)
- You want deeper implementation references
What Information These Files Contain
The APIpie LLM context files describe the platform at the system level, including:
- Core purpose of APIpie (unified AI API, routing, aggregation)
- Supported providers and model families
- Routing, fallback, and observability concepts
- High-level API structure and entry points
- Links to authoritative documentation and MCP endpoints
They do not include UI details, marketing language, or tutorials.
Usage with AI Tools
Cursor
Cursor can use APIpie’s LLMs.txt files to provide better assistance when working with AI routing, provider selection, and API integration.
How to use:
- Reference the APIpie LLMs.txt URLs in chat
- Add them to your project context using
@docs
Example:
@docs https://apipie.ai/llms.txt
Read more about Cursor Web and Docs Search
Windsurf
Windsurf can load APIpie’s LLM context files to understand routing behavior and platform APIs.
Usage:
- Reference
@docswith the APIpie LLMs.txt URLs - Optionally create persistent workspace rules pointing to these files
Read more about Windsurf Web and Docs Search
ChatGPT, Claude, and Other LLMs
Any LLM that supports external documentation references can use APIpie’s LLM files.
Examples:
- “Using APIpie documentation from https://apipie.ai/llms.txt”
- “Follow the full APIpie routing and provider guidelines from https://apipie.ai/llms-full.txt”
These references help the model stay aligned with current production behavior, not training-time assumptions.
Relationship to MCP
LLMs.txt files and MCP serve different purposes:
| Capability | LLMs.txt | MCP |
|---|---|---|
| Purpose | Platform understanding | Live interaction |
| Access | Static, preloaded | Dynamic, tool-based |
| Cost | Zero runtime cost | Per request |
| Use case | Discovery, planning | Querying, execution |
LLMs.txt answers “what is APIpie and how does it work?”MCP answers “interact with APIpie right now.”
Both are intentionally provided.
Stability and Guarantees
APIpie treats its LLMs.txt files as public contracts:
- Routes are stable
- Content reflects production behavior
- Changes are intentional and reviewed
- AI systems are expected to rely on them safely
If an AI tool misunderstands APIpie, this is the first place that gets fixed.