One JSON block connects any MCP agent.
Add Bilinc as an MCP server. Restart the agent. Tools appear automatically. No SDK imports, no HTTP wiring, no polling loops.
Bilinc exposes commit, recall, and status as MCP tools. Any agent that speaks MCP gets durable, verifiable memory without writing HTTP clients.
Add Bilinc as an MCP server. Restart the agent. Tools appear automatically. No SDK imports, no HTTP wiring, no polling loops.
{
"mcpServers": {
"bilinc": {
"command": "python",
"args": ["-m", "bilinc.cloud_mcp"],
"env": {
"BILINC_API_KEY": "bil_live_..."
}
}
}
}MCP (Model Context Protocol) is an open standard for connecting AI agents to external tools and data sources. When configured as an MCP memory server, Bilinc lets any MCP-supporting agent commit, recall, and manage durable memory through standardized tool calls instead of raw API calls.
Add a bilinc entry to your agent's MCP configuration. For Claude Code or Codex, add to the agent's config file: { "mcpServers": { "bilinc": { "command": "python", "args": ["-m", "bilinc.cloud_mcp"], "env": { "BILINC_API_KEY": "$YOUR_KEY" } } } }. Then restart the agent and tools like commit_mem and recall become available.
Claude Code, OpenAI Codex CLI, Cursor, Windsurf, and any agent that implements the MCP client specification can use Bilinc as a memory server. The MCP adapter exposes commit_mem, recall, status, and health tools that the agent discovers at startup.
Yes, the current MCP adapter (python -m bilinc.cloud_mcp) connects to Bilinc Cloud. You need a BILINC_API_KEY from the Cloud dashboard. Start with a free 7-day trial: bilinc start, then bilinc login --api-key <key>.
The MCP adapter provides: commit_mem (store typed semantic, episodic, or procedural memory with importance), recall (retrieve memories by query, type, or key), status (show connection health and memory counts), and health (ping the Cloud endpoint).