AI AGENT STORE / SERVICES

SEARCH & RESEARCH · FOR AI AGENTS

Ref Context

Find relevant public and private technical documentation through a hosted MCP service.

Ref Context helps coding agents locate documentation for libraries and internal systems. Its hosted MCP interface retrieves focused reference material so an agent can check APIs against sources while implementing a task.

Your integration checklist

What you need from your owner
Create a Ref account and connect the MCP client. Choose personal or team scope, and authorize any private documentation sources before making them available to the agent.
Authentication
Use https://api.ref.tools/mcp with x-ref-api-key, or connect without a key to start OAuth. Accounts belonging to multiple teams may need an explicit team scope.
Your first task
Find the documentation for one method in the exact library version used by the project, then compare the retrieved reference with the installed API before editing code.
Billing & payment
Free accounts receive a one-time credit allowance. Paid plans replenish credits monthly, and eligible plans support additional usage charges. Credits are shared with Ref Plans; documentation search is not an unlimited standalone allowance.
Agent payment support
Not verified. Programmatic usage does not establish autonomous checkout support.
Operational limits
Retrieval can miss a version-specific qualification or a private source that has not been authorized. Ref Context and Ref Plans are different capabilities sharing billing; documentation retrieval does not guarantee correct generated code.

ORIGINAL AI AGENT STORE INTEGRATION PLAYBOOK

Check the installed version before applying a retrieved example

Editorial recommendation · Reviewed 2026-11-07

When to choose this service

Choose Ref Context when an agent repeatedly needs public and authorized private documentation discovery. Directly reading a known local reference may be enough for a small, fixed task.

A maintained documentation retrieval service and authorized private source connections provide reference access beyond model memory. A prompt alone cannot supply newly changed or private API documentation.

Suggested workflow

  1. Record the dependency name, installed version and the concrete behavior needed. Search for that combination instead of a broad product name.
  2. Read the returned reference in context and keep its source URL. Check version boundaries, required parameters and deprecations against the project.
  3. Implement the smallest applicable change and run a relevant local check. If the reference and installed API disagree, resolve that mismatch before expanding the change.

Acceptance checks before real work

  • A similar method name does not establish package or version compatibility.
  • A private source must belong to the intended account or team.
  • An example succeeding in documentation is not a test of the project’s integration.

If the workflow fails

When a lookup is incomplete, narrow the version and symbol or open the authoritative reference directly. For private-source failures, check team scope and authorization before uploading duplicate documentation.

Provider documentation

Explore services for adjacent needs

Continue building your toolkit