fix: permissive mode perf, remove Pollinations, enforce policy#16
Merged
seanfromthepast merged 3 commits intomainfrom Mar 12, 2026
Merged
fix: permissive mode perf, remove Pollinations, enforce policy#16seanfromthepast merged 3 commits intomainfrom
seanfromthepast merged 3 commits intomainfrom
Conversation
- Default max_tokens=2048 when model_policy=permissive and no preset - Use batch dialog (1 LLM call) instead of sequential (7 calls) - Skip critique loop in permissive mode (saves 1-8 calls) - Block Google fallback in permissive mode (correctness: stay Google-free) - Reorder model preference to prioritize fast non-thinking models (DeepSeek R1 moved to last resort — its 30-60s/call caused timeouts)
Remove Pollinations.ai as image provider. All image generation now routes through OpenRouter or Google native. Permissive mode uses OpenRouter for images (Google-free). Simplifies fallback chain from 3-tier (Google -> OpenRouter -> Pollinations) to 2-tier (Google -> OpenRouter).
…weight image models - Validate explicit text_model/image_model against PERMISSIVE_PREFIXES when model_policy=permissive. Returns 422 for proprietary models (OpenAI, Anthropic, Google Gemini). - Add is_model_permissive() helper for reuse across validation and registry filtering. - Add permissive_only param to get_best_image_model() and get_image_fallback_model() — skips Gemini/proprietary image models. - Permissive image model selection now prefers open-weight models from the OpenRouter registry instead of defaulting to Gemini.
seanfromthepast
approved these changes
Mar 12, 2026
seanfromthepast
approved these changes
Mar 12, 2026
seanfromthepast
approved these changes
Mar 12, 2026
realityinspector
pushed a commit
that referenced
this pull request
Mar 15, 2026
…cement fix: permissive mode perf, remove Pollinations, enforce policy
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
model_policy="permissive", auto-select open-weight image models via registryTest plan
model_policy=permissivegenerates in ~2min instead of timing out at 600spytest -m fastpasses