The best RAG MCP and CLI for AI agents

Connect Claude or headless agents to your Lookio knowledge base via MCP or CLI. Stop copy-pasting context and let your agents retrieve exactly what they need, when they need it.

Start free

Prompt copied

  1. 1 Open your AI agent
  2. 2 Paste the prompt Ctrl V
  3. 3 Approve the connection when Lookio asks
Show the prompt

Set up Lookio for me: connect to it, import my documents and create an assistant that answers from them. Follow https://lookio.app/agent.md

Claude
A customer asks if their annual plan can be refunded. Draft a reply from our policy.
lookio.ask_assistantAnswered

Hi Dana, annual plans are refunded in full within 30 days of purchase 1, so you are covered. The money goes back to your original payment method 2.

Connect your agent

MCP server, with OAuth or a key

https://mcp.lookio.app/mcp

CLI

npm install --global @lookio/cli
lookio auth login

Works from

ClaudeChatGPTCursor
Direct integration with Claude DesktopUse the Lookio MCP server to let Claude search your proprietary documents, manage resources, and answer questions natively.
Machine-readable JSON via CLIGive headless agents and Python bots access to your knowledge base: the CLI prints JSON whenever its output is piped.
Answers from your documents, not from memoryForce your AI agents to use your verified documents as the single source of truth before they write code or answer queries.
One knowledge base for every agentMaintain a single source of truth in Lookio while exposing it to multiple autonomous agents across your tech stack.

Agents answer better from your own sources

Give agents a controlled path to your real company knowledge.

Sources of this answer
Refund policyPassage 1
Billing helpPassage 2
Also read
Price list 2026
What changes in our plans
Onboarding guide
Warranty terms
Reseller agreement
Shipping FAQ
Battery safety sheet
Refund policyPage 12
01Refunds and cancellations
02
03Monthly plans can be cancelled at any time and stop
04at the end of the current billing period.
05
06Annual plans
07
08Annual plans are refunded in full when the request
09is made within 30 days of purchase.
10After 30 days, the plan runs until the end of the year
11and can be switched to monthly at renewal.
12
13How refunds are paid
14
15A refund always goes back to the original payment
16method and appears within five working days.
17
18Exceptions

Interactive agents with native access

Let compatible agents query Lookio while you work.

Claude CodeTerminal
claude mcp add --transport http lookio https://mcp.lookio.app/mcp

Then /mcp, and approve in your browser

Cursor.cursor/mcp.json
{ "mcpServers": {
  "lookio": {
    "url": "https://mcp.lookio.app/mcp"
  }
} }
ChatGPT and ClaudeConnector

Custom connector URL

https://mcp.lookio.app/mcp

Any client with remote MCP and OAuth

Headless agents and terminal scripts

Bring Lookio retrieval into terminal and script workflows.

Terminal
$ lookio auth login
Signed in to Acme

$ lookio assistants
cf3c3452  Support
a81d07be  Product docs

$ lookio ask cf3c3452 "How long does delivery take?" | jq
stdoutDone, 30 credits
{ "data": {
    "answer": "Orders ship within 2 business days. 1",
    "citations": [{
      "n": 1,
      "title": "Shipping and returns"
    }],
    "credit_cost": 30
} }

Questions about Lookio MCP and CLI

What is the Model Context Protocol (MCP)?

MCP is an open standard that allows AI agents and assistants (like Claude Desktop or Cursor) to securely connect to external tools and data sources. The Lookio MCP Server acts as a bridge, allowing these agents to natively search your Lookio knowledge base instead of relying solely on their base training data.

Why should I use the Lookio MCP Server instead of the standard API?

Use the MCP Server when you are working inside an interactive agent environment (like Cursor or Claude Desktop) and want the agent to autonomously discover and use Lookio. Use the REST API when you are building fixed backend workflows or programmatic automations in tools like n8n or Make.

How does the Lookio CLI help AI coding agents?

The Lookio CLI is optimized for terminal-based and headless agents. Whenever its output is not a terminal, it prints clean, machine-parseable JSON, with no human formatting to strip. This allows Python bots, bash scripts, and CI/CD pipelines to query your knowledge base directly.

Can I connect Cursor to my Lookio workspace?

Yes! By adding the Lookio MCP Server to your Cursor configuration, you allow the Cursor agent to search your internal API guidelines, architectural decisions, and product specs before it generates a single line of code, drastically reducing errors.

Is my knowledge base synced across the API, MCP, and CLI?

Yes. Whether you query via the Lookio web interface, the REST API, the MCP Server, or the CLI, you are accessing the exact same centralized assistants and document resources. Your single source of truth remains consistent everywhere.

Agents answer better from your own sources

Give agents a controlled path to your real company knowledge.

  • Stop feeding 100 PDFs into a prompt.
  • Reduce token usage and costs.
  • Ensure agents use up-to-date specs.

Interactive agents with native access

Let compatible agents query Lookio while you work.

  • Connect via standard MCP.
  • Query assistants inside the chat.
  • Manage resources directly from Claude.

Headless agents and terminal scripts

Bring Lookio retrieval into terminal and script workflows.

  • Zero installation via npx.
  • Returns clean, parseable JSON output.
  • Secure local key management.

Create your first assistant.

Start free

Prompt copied

  1. 1 Open your AI agent
  2. 2 Paste the prompt Ctrl V
  3. 3 Approve the connection when Lookio asks
Show the prompt

Set up Lookio for me: connect to it, import my documents and create an assistant that answers from them. Follow https://lookio.app/agent.md

1,000 free credits, no credit card