Examples

New here? Start with the notebooks

These scripts are the runnable distillation of the interactive notebooks — learn the concept step by step in a notebook, then ship it from examples/.

Runnable examples tested on NVIDIA Jetson AGX Thor (132GB unified memory).

---

Demo Video

Click to watch the full demo video

---

All Examples

• 01 — Basic Text (Physics Reasoning)

  

  Text-only physics reasoning — no video or image needed. ~11s on Thor.

  → Full example + code

• 02 — Video Captioning

  

  Detailed temporal-spatial descriptions from video. ~15s on Thor.

  → Full example + code

• 03 — Driving Analysis (CoT)

  

  Dashcam safety analysis with chain-of-thought reasoning. ~16s on Thor.

  → Full example + code

• 04 — Embodied Reasoning

  

  Robot next-action prediction from workspace images. ~43s on Thor.

  → Full example + code

• 05 — Tool Usage

  

  Cosmos as a callable tool inside any Strands agent. ~9s on Thor.

  → Full example + code

• 06 — Cosmos 3 Reasoner

  Omnimodal video/image understanding via local vLLM — caption, temporal, embodied, grounding.

  → examples/06_cosmos3_reason.py · Cosmos 3 Guide

• 07 — Cosmos 3 Generator

  Text → image / video / video + sound (in-process Diffusers).

  → examples/07_cosmos3_generate.py

• 08 — Cosmos 3 Action

  World-model rollouts: forward/inverse dynamics, policy (Cosmos Framework).

  → examples/08_cosmos3_action.py

• 09 — Cosmos 3 Showcase (Reason → Generate)

  

  Reason about a real video, then generate similar videos (incl. audio) from the description.

  → examples/09_cosmos3_showcase.py · demo/cosmos3_showcase/

---

Quick Reference

#	Example	Time (Thor)	Recording
1	Basic Text	~11s	cast
2	Video Caption	~15s	cast
3	Driving Analysis	~16s	cast
4	Embodied Reasoning	~43s	cast
5	Tool Usage	~9s	cast
6	Cosmos 3 Reasoner (vLLM)	caption ~5s	—
7	Cosmos 3 Generator (Diffusers)	t2v ~20–55s	—
8	Cosmos 3 Action (Framework)	rollout ~30s	—
9	Cosmos 3 Showcase (reason→generate)	full loop	—

---

Running Locally

git clone https://github.com/cagataycali/strands-cosmos.git
cd strands-cosmos
pip install -e .

# Jetson devices: fix CUBLAS first
strands-cosmos-fix-cublas

# Run any example
python examples/01_basic_text.py
python examples/02_video_caption.py
python examples/03_driving_analysis.py
python examples/04_embodied_reasoning.py
python examples/05_tool_usage.py

# Cosmos 3 (see the Cosmos 3 Guide for env setup: just c3-setup-reason / c3-setup-gen)
python examples/06_cosmos3_reason.py       # needs `just c3-serve-reason` running
python examples/07_cosmos3_generate.py     # needs `just c3-setup-gen`
python examples/09_cosmos3_showcase.py     # reason -> generate showcase

Sample media

Examples 02–05 need a sample.mp4 (video) and/or sample.png (image) in the project root. Set paths via environment variables:

export SAMPLE_VIDEO=/path/to/your/video.mp4
export SAMPLE_IMAGE=/path/to/your/image.png

Playing Terminal Recordings

All examples have asciinema .cast recordings:

pip install asciinema

# Play any recording
asciinema play docs/assets/casts/01_basic_text.cast
asciinema play docs/assets/casts/03_driving_analysis.cast

---

Execution Flow

graph TD
    START["Run Example"] --> MODEL["Load Model<br/>~3s (cached)"]
    MODEL --> MEDIA{"Has media?"}
    MEDIA -->|"Video"| DECODE["Decode frames<br/>@ configured FPS"]
    MEDIA -->|"Image"| PROCESS["Process image<br/>visual tokens"]
    MEDIA -->|"Text only"| TOKENIZE["Tokenize text"]
    DECODE --> INFER["GPU Inference<br/>token-by-token streaming"]
    PROCESS --> INFER
    TOKENIZE --> INFER
    INFER --> OUTPUT["Stream output<br/>to terminal"]
    OUTPUT --> DONE["✅ PASS"]

    style MODEL fill:#264653,color:#fff
    style INFER fill:#76b900,color:#fff
    style DONE fill:#2d6a4f,color:#fff