Picteus: open-source local library keeps every AI image with its full generation recipe
shivam_dewan · reddit · 2026-10-14
Developer shivamdewan is building Picteus, a free open-source app that runs locally and keeps every generated image together with its recipe — prompt, checkpoint, LoRAs and weights, seed, workflow — so creators stop losing how their best work was made.
Current features:
- Search by recipe: find images by prompt words, checkpoint, or LoRA name.
- Search by content: CLIP and DINO run locally for description-based or image-similarity search.
- Local captioning: tags and captions via Ollama or llama.cpp, nothing uploaded.
- Extension architecture: ComfyUI, Midjourney, CivitAI are all extensions; write your own, even with an AI coding assistant.
- In-app generation: Gemini and Replicate work via extensions; generation using your library as context is next.
The author admits first-time setup is rough and is asking users for 5-minute feedback calls.
More from Multimodal
- Gemini 4 Argon recreates Manhattan photo as oil painting via multi-pass canvas renderer — _philschmid · 2026-10-14
- Hollywood embraces AI tools despite doomsday fears, Variety reports from studio stages — DavidmComfort · 2026-10-14
- Early user teases Gemini 4 Argon: impressive 3D modeling and spatial reasoning — JMateosGarcia · 2026-10-14
- Netflix's Eyeline Studios details its AI 'hybrid production' pipeline: relighting, location swaps and camera repositioning after the shoot — DavidmComfort · 2026-10-14
- fal launches Reframe: change any video's aspect ratio without cropping or reshooting — jfischoff · 2026-10-14
- CineConIA builds free ComfyUI nodes for AI filmmaking camera control — CineConIA · 2026-10-14