Integration

CodeGPT integration

CodeGPT (codegpt.co) hosts AI agents that answer from a knowledge base of uploaded documents. List, create, update and delete agents, attach documents to an agent, chat with an agent, upload, read and delete knowledge-base documents, and generate or edit a document's metadata. The API is a beta that CodeGPT says may change. Every chat call counts as an interaction billed by the plan (API plan: $20 per 1,000 interactions, $50 monthly minimum, plus model token use). Agent sharing settings (public flag, PIN code) and account profile edits are not exposed. Authentication is a CodeGPT API key sent as a Bearer token; pass the optional CodeGPT-Org-Id header to work inside an organization.

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in CodeGPT.Built from the 16 CodeGPT tools below, on the connection you add once.

What you can do with CodeGPT

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Agent, 5 operations

  • create_agentCreate agentPOSTCreates an agent with a name and optional model, prompt, temperature, topk and welcome message. Returns the new agent's fields as CodeGPT sends them back.
  • delete_agentDelete agentDELETEDeletes an agent for real. Its chats stop working.
  • get_agentGet agentGETReturns one agent. Set populate to agent_documents to also get the ids of the documents in its knowledge base.
  • list_agentsList agentsGETLists the agents of the account (or of the organization in CodeGPT-Org-Id) with their model, prompt, temperature and settings.
  • update_agentUpdate agentPATCHUpdates an agent's name, model, temperature, prompt, topk or welcome message. Only the fields sent change; leave a field out to keep it. Changes the live agent.

Document, 5 operations

  • delete_documentDelete documentDELETEDeletes a knowledge-base document for real; agents that used it lose that knowledge.
  • get_documentGet documentGETReturns one knowledge-base document, including its full text content and metadata.
  • list_documentsList documentsGETLists the knowledge-base documents with their file type, token and chunk counts and metadata (title, summary, keywords, language).
  • set_agent_documentsSet agent documentsPATCHSets the agent's knowledge-base documents to exactly the list sent, replacing the current list. Read it first with get_agent (populate=agent_documents).
  • upload_documentUpload documentPOSTUploads a file into the knowledge base and returns its documentId. Attach it to an agent with set_agent_documents.

Metadata, 2 operations

  • generate_document_metadataGenerate document metadataPOSTGenerates metadata (title, description, summary, keywords, language) for an uploaded .txt file without storing it. Read-only.
  • update_document_metadataUpdate document metadataPATCHReplaces a document's metadata. title, summary, keywords and language are required: read the current values with get_document and send them back with the change.

Completion, 1 operation

  • chat_with_agentChat with agentPOSTSends messages to a CodeGPT agent and returns its reply in choices[0].message.content. Each call counts as one billed interaction plus model tokens.

Image, 1 operation

  • update_agent_imageUpdate agent imagePATCHReplaces the agent's chat image with an uploaded jpeg, png or webp file (5 MB max).

Marketplace favorite, 1 operation

  • list_favorite_marketplace_agentsList favorite marketplace agentsGETLists the marketplace agents the user marked as favorites.

User, 1 operation

  • get_current_userGet current userGETReturns the user the API key belongs to: username, full name, avatar URL and organization ids.

Use CodeGPT from Claude Code, Cursor or Codex

Connect CodeGPT in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its CodeGPT operations become tools the client can call, next to every other integration you connected.

Claude Code
claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"
Codex (~/.codex/config.toml)
[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"
Cursor, Claude Desktop and other JSON configs
{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect CodeGPT

  1. Create your workspace

    Sign up free. Nothing to install.

  2. Connect CodeGPT once

    Add your CodeGPT connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.

  3. Describe the job in chat

    The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.

CodeGPT integration FAQ

What can you do with the CodeGPT integration?

LiveContext exposes 16 operations on CodeGPT, covering agent, document, metadata and completion.

Popular integrations

Browse the whole catalog