Skip to main content

MCP Server

OpenTryOn ships a Model Context Protocol server under mcp-server/. Every model in tryon.cli.registry becomes an MCP tool automatically — the same surface as the opentryon CLI, via tryon.cli.runner.invoke_model().

Current release: works with OpenTryOn v0.0.3+ (pip install -U opentryon).

Why it matters​

  • Agents in Cursor, Claude Desktop, or tryon-studio can call try-on, generate, edit, video, understand, and bg-remove tools directly
  • New registry models appear as tools with zero hand-written MCP wrappers
  • CLI and MCP cannot drift — one runner, one registry

Quick start​

cd mcp-server
pip install -r requirements.txt   # includes fastmcp
cp ../env.template ../.env        # fill API keys you need
python server.py                  # stdio (default for most MCP clients)
# or:
python server.py --transport http --host 127.0.0.1 --port 8000

Discovery tools:

  • list_opentryon_tools — list services / models / tool names / env readiness
  • opentryon_status — configuration status report

Every generated tool accepts dry_run and output_dir, matching the CLI.

Client config​

See example configs in the repo:

Full tool tables and architecture notes: mcp-server/README.md.