saccade good first issues (Python)

whisper.cpp ASR backend (Apple Silicon / Metal acceleration) 14 days ago

enhancement help wanted priority: high area: asr area: performance

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Semantic search: optional multilingual embeddings with hybrid ranking 14 days ago

enhancement help wanted priority: high area: search

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Re-run the benchmarks including frame extraction, and add a frames column 17 days ago

documentation good first issue area: performance

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Measure test coverage in CI 17 days ago

good first issue area: ci

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Docs: recipes for OpenAI-compatible providers (OpenRouter, LM Studio, vLLM, Gemini, Groq) 17 days ago

documentation good first issue area: llm

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Add `Answer.cost()` to estimate the price of an answer 17 days ago

enhancement good first issue area: llm

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

More minimal examples: search-only (offline), OpenAI, Ollama 17 days ago

documentation good first issue

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

CLI: expose ask() options (--no-explore, --max-steps, --max-views, --reasoning-effort) 17 days ago

enhancement good first issue area: llm area: cli

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Export transcripts as Markdown (`Transcript.to_markdown()`, `-f md`) 17 days ago

enhancement good first issue area: cli

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

CLI: a single-line progress bar on interactive terminals 17 days ago

enhancement good first issue area: cli

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Add Python 3.14 to CI and the package classifiers 17 days ago

good first issue area: ci

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Documentation website (MkDocs Material + API reference) on GitHub Pages 17 days ago

documentation help wanted

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Stream answers from ask() as they are generated 17 days ago

enhancement help wanted area: llm area: cli

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Native Anthropic (Claude) client with images, tool use and prompt caching 17 days ago

enhancement help wanted area: llm

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Enable video exploration with local Ollama models that support tools and vision 17 days ago

enhancement help wanted area: llm

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Handle videos that switch language (code-switching) 17 days ago

enhancement help wanted area: asr

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Search and ask across many videos (collections) 17 days ago

enhancement help wanted area: search area: llm

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Cache management: `saccade cache list / prune / clear` 17 days ago

enhancement help wanted area: cli

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Accuracy benchmarks on real recordings (WER, timestamp error, ask() quality) 17 days ago

help wanted priority: high area: performance

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Cross-platform integration tests with real models, run in CI 17 days ago

help wanted priority: high area: ci

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

Speaker diarization: who said what 17 days ago

enhancement help wanted priority: high area: asr

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

On-screen text (OCR) for frames, searchable like speech 17 days ago

enhancement help wanted priority: high area: frames area: search

gabe-santana/saccade

1

Fast, local video context for LLMs: transcripts, frames, timestamped search and ask(), on CPU or GPU.

Python

#llm#rag#video-processing

gabe-santana/saccade

Open Issues Need Help