| --- |
| title: biopython MCP Service |
| emoji: 🔧 |
| colorFrom: blue |
| colorTo: indigo |
| sdk: docker |
| pinned: false |
| license: mit |
| --- |
| |
| # biopython MCP Service |
|
|
| This deployment package exposes core Biopython sequence-analysis capabilities as an MCP service using FastMCP. |
|
|
| ## Available Tools |
|
|
| - `health_check` |
| - `parse_fasta_text` |
| - `translate_dna` |
| - `reverse_complement` |
| - `global_align` |
| - `compute_gc_fraction` |
| - `find_motif_positions` |
|
|
| ## Local stdio (Claude Desktop / CLI) |
|
|
| ```bash |
| cd biopython/mcp_output |
| MCP_TRANSPORT=stdio python start_mcp.py |
| ``` |
|
|
| You can also run the local stdio entry directly: |
|
|
| ```bash |
| cd biopython/mcp_output/mcp_plugin |
| python main.py |
| ``` |
|
|
| ## HTTP MCP (Docker / HF Spaces) |
|
|
| This deployment uses FastMCP HTTP transport, with endpoint exposed at `/mcp`. |
|
|
| Local run: |
|
|
| ```bash |
| MCP_TRANSPORT=http MCP_PORT=7860 python biopython/mcp_output/start_mcp.py |
| ``` |
|
|
| Docker run: |
|
|
| ```bash |
| ./run_docker.sh |
| ``` |
|
|
| Then connect MCP clients to: |
|
|
| - `http://localhost:7860/mcp` |
| - `https://<your-space-host>/mcp` |
|
|
