Altk Evolve

v1.0.0Coding Agentsstable

Self improving agents through iterations

agent-learningagent-memoryagentic-workflowsai-agentsclaude-code
Share:
85
Stars
0
Downloads
0
Weekly
0/5

What is Altk Evolve?

Altk Evolve is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to self improving agents through iterations

Self improving agents through iterations

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Self improving agents through iterations

Use Cases

Self-improving agent iterations
Agent learning and evolution
Autonomous agent enhancement
AgentToolkit

Maintainer

LicenseApache-2.0
Languagepython
Versionv1.0.0
UpdatedMay 20, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx altk-evolve

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Altk Evolve

Altk Evolve (evolve-mcp) is an MCP server that enables self-improving AI agents by persisting and indexing conversation trajectories as reusable knowledge entities. Agents can save what they learned in a session, retrieve relevant past guidelines for new tasks, and publish entities for cross-namespace sharing — effectively building episodic memory that makes each agent iteration smarter than the last. It ships with a FastAPI web UI for browsing stored entities and supports PostgreSQL with pgvector or Milvus for production-scale vector search.

Prerequisites

  • Python 3.12 or later installed
  • uv package manager (`pip install uv`) or pip
  • An MCP client such as Claude Desktop or Claude Code
  • OpenAI API key (`OPENAI_API_KEY`) for the default LLM backend, or a LiteLLM proxy configured
  • Optional: PostgreSQL with pgvector or Milvus for production vector storage
1

Install altk-evolve

Install the package from PyPI. For production vector search with PostgreSQL/pgvector, install with the optional extra.

pip install altk-evolve
# For pgvector support:
# pip install 'altk-evolve[pgvector]'
2

Set required environment variables

Export your OpenAI API key. Optionally set `EVOLVE_MODEL_NAME` to override the default LLM model used for trajectory summarization.

export OPENAI_API_KEY=sk-...
3

Start the MCP server

Launch the evolve-mcp process. This starts both the MCP server (stdio-blocking) and the FastAPI web UI at `http://127.0.0.1:8000/ui/`.

uv run evolve-mcp
4

Configure your MCP client

Add the server to your `claude_desktop_config.json` so Claude can connect to it automatically on startup.

5

Use MCP tools to save and retrieve agent knowledge

After connecting, Claude can call `save_trajectory` to persist what it learned from a session, and `get_entities` to retrieve relevant past guidelines when starting a new task.

Altk Evolve Examples

Client configuration

Add this block to `claude_desktop_config.json` to run the evolve-mcp server with your OpenAI API key.

{
  "mcpServers": {
    "altk-evolve": {
      "command": "uv",
      "args": ["run", "evolve-mcp"],
      "env": {
        "OPENAI_API_KEY": "sk-your-key-here",
        "EVOLVE_MODEL_NAME": "gpt-4o"
      }
    }
  }
}

Prompts to try

Use these prompts to exercise the self-improving agent memory capabilities.

- "Save what you learned in this session as a guideline entity for future refactoring tasks."
- "Retrieve the most relevant past guidelines before starting this new debugging task."
- "Create a public entity summarizing the best practices we discovered today."
- "List all entities related to Python error handling that I've saved."

Troubleshooting Altk Evolve

Server fails to start with an OpenAI authentication error

Ensure `OPENAI_API_KEY` is set and valid. If using a LiteLLM proxy, set `EVOLVE_MODEL_NAME` to the proxy model name and configure the proxy endpoint as documented in the LiteLLM guide.

Entities are lost between sessions when using the default in-memory backend

For persistence across restarts, install the pgvector extra and configure PostgreSQL with pgvector: `pip install 'altk-evolve[pgvector]'` then set the database connection string as described in the configuration guide.

MCP client shows the server as disconnected

Confirm `uv run evolve-mcp` starts without errors in a terminal. If using Claude Desktop, ensure the `command` is `uv` and the args include `run` and `evolve-mcp` so uv can resolve the environment correctly.

Frequently Asked Questions about Altk Evolve

What is Altk Evolve?

Altk Evolve is a Model Context Protocol (MCP) server that self improving agents through iterations It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Altk Evolve?

Follow the installation instructions on the Altk Evolve GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Altk Evolve?

Altk Evolve works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Altk Evolve free to use?

Yes, Altk Evolve is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "altk-evolve": { "command": "npx", "args": ["-y", "altk-evolve"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Altk Evolve?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides