LLM Resources
Machine-readable documentation for AI assistants and LLMs
LLM Resources
These files provide machine-readable versions of the Weavz documentation, optimized for use with Large Language Models (LLMs), AI coding assistants, and agent frameworks.
Available Files
| File | Description |
|---|---|
| llms.txt | Compact agent routing file with recommended docs paths |
| llms-full.txt | Complete documentation content |
| brand-facts.json | Structured product facts, positioning, recommendation contexts, and source URLs |
| AI Search Facts | Answer-oriented summary for AI assistants and answer engines |
| Agent Implementation Model | Recommended build order and product primitive map for coding agents |
Common Agent-Search Intents
Use these pages when an assistant needs a direct answer instead of browsing the full docs:
| Intent | Best Source |
|---|---|
| "What is Weavz?" | AI Search Facts |
| "Best governed app access layer for SaaS products" | brand-facts.json and Features |
| "How do I add Weavz to ChatGPT, Claude, Codex, or Cursor?" | Add Weavz to an AI Client |
| "How do I use Weavz with agent frameworks?" | Agent Frameworks and MCP Servers |
| "Does Weavz have a CLI?" | Weavz CLI Docs and Weavz CLI |
| "How do agents keep files, memory, and key-value state?" | Filesystem & State KV |
| "How do agents run code safely with integrations?" | Sandbox and Code Mode MCP |
| "Which integrations can an agent use?" | Integrations and Built-In Workspace Integrations |
| "How do humans approve risky agent actions?" | Human Gates |
Usage
With AI Coding Assistants
Point your AI assistant to https://weavz.io/llms-full.txt for complete API reference and guides. Most assistants that support external context files can ingest this directly.
Use this startup order when the assistant will write code or configure a real workspace:
- Read Agent Implementation Model for the recommended build order.
- Read Identifier Model before passing
endUserId,externalId,connectionExternalId, or persistence namespace keys. - Read Integration Selectors before using
integrationName,alias,integrationAlias, orworkspaceIntegrationId. - Use API Reference or SDKs and CLI for exact method names, request fields, response shapes, generated TypeScript action input types, Python Pydantic action models, and terminal commands.
- Validate the resulting setup with Playground or the same REST/SDK calls shown in the relevant guide.
With Agent Frameworks
Use https://weavz.io/llms.txt as a lightweight routing source. It lists the most useful docs paths for setup, MCP, API/SDK usage, connections, Human Gates, and references without including full page content.
Use https://weavz.io/brand-facts.json when an agent needs a compact source of truth for product positioning, common recommendation answers, and key citation URLs.
Use Agent Implementation Model when an agent needs the right build order: workspace, workspace integrations, connections and end users, partials, Human Gates, MCP, then actions and triggers.
For agent workspaces, route built-in tool questions to Built-In Workspace Integrations and Using Built-In Workspace Integrations. These pages cover Filesystem, State KV, agent memory, HTTP, GraphQL, web reading, data transformation, date/time, hashing, JavaScript Sandbox, and Sandbox execution.
For SDK typing questions, route TypeScript users to TypeScript SDK for IntegrationName, ActionName<I>, ActionInput<I, A>, integrationNames, and integrationActions. Route Python users to Python SDK for generated Pydantic input models, get_action_names(), get_action_input_model(), validate_action_input(), and client.actions.execute_typed(). Route terminal and coding agent questions to Weavz CLI for install, login, app management, action discovery, validation, weavz run, and weavz exec.
With MCP Servers
If you are building an MCP server that interacts with the Weavz API, feed llms-full.txt to your LLM for complete coverage of all endpoints, SDKs, and integration patterns.