Record extended Telegram requirements for profile-driven RunPod UX

This commit is contained in:
kflame99
2026-02-15 09:38:16 +00:00
parent c4953c69b7
commit fe41179fd1
2 changed files with 14 additions and 0 deletions
+6
View File
@@ -12,6 +12,12 @@
- `Runpod_Comfy/config/custom_nodes.lock.json`
- `Runpod_Comfy/config/models.json`
- Docker base image is taken from generated `comfyui.lock` (`BASE_IMAGE` build arg).
- Goal: normal updates (new node repo/model URL/version bumps) should require editing config files only, not Python/Docker logic.
- Expected operator UX:
- Edit 3–4 simple profile files with real values.
- Run compiler.
- Build/push image.
- Deploy and smoke-test on RunPod.
## 2026-02-15 — Work continuity model
- Planning/serious execution in OpenClaw web dashboard.
+8
View File
@@ -34,3 +34,11 @@
### Blockers
- Real model URLs/checksums not finalized.
### Additional context imported from Telegram (extended)
- Desired abstraction confirmed: config-driven workflow where adding node repos/model URLs and bumping versions does not require code edits.
- Strong emphasis on "production-ready" compiler/scaffold with placeholders now and real values tomorrow.
- Tomorrow checklist reaffirmed:
1. Fill real profile values (CUDA/image/nodes/models).
2. Compile profile, build image, push image.
3. Deploy on RunPod and validate.