Open-source infrastructure that reduces AI agent token consumption by 70-98%.
Your AI writes code in a secure nsjail sandbox instead of loading tool schemas into the context window. Data stays in the sandbox, never enters context. Fewer tokens, same results.
Self-host the full platform free forever. Released under BSL 1.1 — converts irrevocably to Apache 2.0 in 2030.
| Repository | Description | License |
|---|---|---|
| mcpworks-api | Core platform — FastAPI backend, nsjail code sandbox, agent runtime, namespace hosting | BSL 1.1 |
| linkedin-pages-mcp | MCP server for LinkedIn Company Page management via the official Community Management API | MIT |
git clone https://github.com/MCPWorks-Technologies-Inc/mcpworks-api.git
cd mcpworks-api
cp .env.self-hosted.example .env
docker compose -f docker-compose.self-hosted.yml up -dAdd to your AI assistant's .mcp.json:
{
"mcpServers": {
"myns-create": {
"type": "http",
"url": "https://api.mcpworks.io/mcp/create/myns",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
}
}
}Built in Vancouver, Canada by MCPWorks Technologies Inc.