Detect

1. Confirm Python project — at least one of:
   • requirements.txt
   • pyproject.toml
   • *.py files in workspace root
2. Scan dependencies in requirements.txt or pyproject.toml:

3. Locate the WSGI / ASGI entry point (best-effort):
   • Flask common: app.py exporting app, application.py, wsgi.py
   • Django common: <project>/wsgi.py
   • FastAPI common: main.py exporting app
4. Record findings in your working memory so Step 5 (startup) can use them.

Outcomes

Priority rule when both flask and fastapi appear in requirements.txt (or pyproject.toml): always treat as FastAPI — set the explicit uvicorn startup command. This is deterministic and avoids relying on import-order or token-order heuristics. Rationale: Flask is happily auto-detected by Oryx with no startup command, but FastAPI requires the explicit uvicorn command to run reliably; if the project actually uses Flask as the served app, the user can override the startup command later, but if it uses FastAPI and we silently picked Flask, the container ping fails.

Important rules

• ⛔ Do not abort deployment when framework is unknown.
• ⛔ Do not try to install Flask, Django, or any framework into the user’s project.
• ✅ Always report what was detected so the user has full context.
• ✅ When detection is wsgi-generic, asgi-generic, or unknown, remember this fact for Step 8 — the post-deploy message must use the unknown-framework template in post-deploy-message.md, which adds an explicit “set a startup command” instruction. The Step 5 warning alone is not enough; the user needs the reminder at the end of the run too.

Example output to surface to the user

or

or

or