nemostation 0.1.17

Find moments in your videos. Agent-first CLI for Marlin-2B: dense captioning + temporal grounding, local or hosted.

The command-line tool for Marlin-2B — a 2B video VLM for the two questions you actually ask a video: what is happening, and when. Runs free and local on Apple Silicon — no API key, no Hugging Face account.

• marlin caption → a Scene description + a <start>–<end> event timeline
• marlin find → the single start → end span where your query happens

Install

Apple Silicon (M-series Mac) only for now. NVIDIA / other platforms are coming as a separate optimized build.

uv tool install nemostation      # or: pipx install nemostation
marlin setup                     # one-time: sign in, build the local engine
marlin caption clip.mp4          # describe what's in a video
marlin find clip.mp4 "a deer crossing"   # locate when it happens → start → end

First run opens your browser for a quick sign-in (two questions, then Google), detects your Mac, and builds the local MLX engine. The 8-bit weights are public — nothing gated. Add --json to any command for parseable output.

What it produces

marlin caption "video.mp4" — what's in it	marlin find "video.mp4", "gunfight" — when it happens

Each call runs one model pass on one bounded clip (~2 min at 2 fps) — the same contract as the inference server. For longer videos, window with ffmpeg and loop.

Why Marlin

At 2B params it's the strongest open model in its weight class on dense captioning (DREAM-1K, CaReBench) and natural-language temporal grounding (TimeLens-Bench) — competitive with Gemini-2.5 at a fraction of the cost. See the benchmarks on the model card.

Use it from an agent

marlin skills install        # → .claude/skills/ + .agents/skills/

Installs the video-understanding skill so Claude Code / Codex use marlin as "eyes on a video" — clip-length and single-find limits baked in. Every verb honors --json (stdout parseable, progress on stderr).

Contributing

Marlin is meant to be extended — and adding a skill is the easiest way in. A skill is a folder under skills/ with a SKILL.md that teaches an agent to use caption / find for one job — clip scoring, b-roll search, highlight reels, footage catalogs, whatever you build.

skills/
  video-understanding/SKILL.md   # ships today — the reference
  your-skill/SKILL.md            # ← add yours

Add one:

1. Copy the format from video-understanding/SKILL.md — frontmatter (name, description, requires.bins) + a short recipe.
2. Keep it honest about the limits (one bounded clip per call; find returns one span).
3. Open a PR. New skill ideas, issues, and docs fixes are all welcome too.

Hack on the CLI:

git clone https://github.com/shu-bamma/marlin-cli
cd marlin-cli
uv tool install --editable .     # or: pip install -e .
pytest                           # contract tests

New verbs, engine support, and bug fixes are all fair game — open an issue to chat about anything bigger. Licensed under Apache-2.0.

Links

• Try it live → vlm.nemostation.com
• Model card + benchmarks → huggingface.co/NemoStation/Marlin-2B
• Team / custom fine-tuning → nemostation.com