- Cactus (YC S25) is a low-latency engine for mobile devices & wearables.
- Cactus runs locally on edge devices with hybrid routing of complex tasks to cloud models like Gemini.
- Google DeepMind just released Gemma 4, the first on-device model you can voice-prompt.
- Gemma 4 on Cactus is multimodal, supporting voice, vision, function calling, transcription and more!
- All teams MUST build products that use Gemma 4 on Cactus.
- All products MUST leverage voice functionality in some way.
- All submissions MUST be working MVPs capable of venture backing.
- Winner takes all: Guaranteed YC Interview + GCP Credits.
- Best On-Device Enterprise Agent (B2B): Highest commercial viability for offline tools.
- Ultimate Consumer Voice Experience (B2C): Best use of low-latency compute to create ultra-natural, instantaneous voice interaction.
- Deepest Technical Integration: Pushing the boundaries of the hardware/software stack (e.g., novel routing, multi-agent on-device setups, extreme power optimization).
Prizes per special track:
- 1st Place: $2,000 in GCP credits
- 2nd Place: $1,000 in GCP credits
- 3rd Place: $500 in GCP credits
- Rubric 1: The relevnance and realness of the problem and appeal to enterprises and VCs.
- Rubric 2: Correcness & quality of the MVP and demo.
- Step 1: Fork this repo, clone to your Mac, open terminal.
- Step 2:
git clone https://github.com/cactus-compute/cactus - Step 3:
cd cactus && source ./setup && cd ..(re-run in new terminal) - Step 4:
cactus build --python - Step 5:
cactus download google/functiongemma-270m-it --reconvert - Step 6: Get cactus key from the cactus website
- Sept 7: Run
cactus authand enter your token when prompted. - Step 8:
pip install google-genai(if using cloud fallback) - Step 9: Obtain Gemini API key from Google AI Studio (if using cloud fallback)
- Step 10:
export GEMINI_API_KEY="your-key"(if using cloud fallback)
- Read Cactus docs carefully: Link
- Read Gemma 4 on Cactus walkthrough carefully: Link
- Cactus & DeepMind team would be available on-site.
- Record a 60s demo video
- Submnit here by 11am
- NO multiple submissions!
