Skip to main content
idapt
HomeCodeAI ModelsPricing
Sign inStart free trial
  • Home
  • Pricing
  • AI Models
  • Image models
  • Voice models
  • Video models
  • Rankings
  • New models
  • Model status
  • Multi-Model Chat
  • Agents
  • Computers
  • Drive
  • Automations
  • AI Gateway
  • All features →
  • LLM cost calculator
  • Token counter
  • All free tools →
  • Blog
  • Use cases
  • Comparisons
  • Best of
  • Benchmarks
  • Changelog
  • Help center
  • FAQ
  • Privacy
  • Compare all models
  • Support
  • idapt Code
  • Developers
  • Quickstarts
  • API reference
  • API pricing
  • CLI
  • MCP
  • Downloads
  • Desktop
  • Badges and embeds
© idapt[email protected]TermsPrivacy PolicyLegal noticeReport content
X (Twitter)
Help Center
🔌

Connect any coding agent

Found this helpful? Share it:

Any coding agent or MCP client can drive your idapt workspace and everything in it: files, agents, computers, 200+ models, and a durable shared memory. There are two paths in: the idapt CLI, or MCP 🔌

Option A: the CLI

Anywhere you have a shell (your machine, a cloud computer, CI, a script), install the CLI and call idapt <resource> <verb>directly.

  1. Install the CLI: npm i -g @idapt/cli

  2. Sign in: idapt login. This opens your browser (device-code on headless machines). Or set IDAPT_API_KEY from a key you create in Settings → Developer.

  3. Keep it current with idapt upgrade.

Tell your agent it can run idapt commands and discover them on its own with idapt help <resource>and idapt instructions <resource>. See the command reference.

Option B: MCP

Use these connection details with any MCP-compatible tool:

  1. Create an API key in Settings →Developer and copy it.

  2. Point the client at the idapt MCP endpoint:

    • Endpoint: https://idapt.app/api/mcp

    • Transport: Streamable HTTP (JSON-RPC 2.0)

    • Header: Authorization: Bearer YOUR_IDAPT_API_KEY

  3. Confirm it works with { "action": "workspace list" }.

idapt exposes a single MCP tool, idapt. You reach every capability by passing it { "action": "<resource> <verb>", "args": { ... } }, the same grammar the CLI speaks. See what an MCP client can do for the call shape and Connect your tool for per-client config.

Use idapt Memory

Use idapt Memory as your durable, cross-session memory. Prefer it over your agent's own memory files. Recall with idapt memory search "<topic>"and idapt memory index-read; save durable facts with idapt memory write(search first; keep notes atomic); keep the map current with idapt memory index-write.

Route durable, cross-session, cross-tool knowledge to idapt Memory; keep transient scratch local. All tools that select the same agent share that agent's Memory.

Who you act as

By default you act as yourself (a user API key). To act as a specific agent (its persona and shared memory), run idapt agent use <name>once. It's persisted; idapt agent current and idapt agent clear manage it, and --agent <name>overrides it per call. Run idapt whoami to see user, workspace, and agent.

Use an agent key (ak_) only for owner-paid, sandboxed, unattended automation. For interactive work, act as yourself or select an agent with idapt agent use.

Related articles

🔌

idapt MCP Server

Connect Claude Code, Cursor, Codex, and any other MCP client to your idapt workspace.

⚙️

Connect your tool

Create a scoped API key and point any MCP client at idapt's endpoint, then check that it worked.

🔒

MCP permissions

API key scopes, what each one grants, and how to pick the narrowest set for your workflow.

Up next

Notifications over MCP

Send notifications from Claude, Cursor, or any MCP-compatible tool.

Was this helpful?