Zerogpu
Offload classification, summarization, NER, PII detection, and short chat to ZeroGPU nano models so your agent stack stays fast and cheap.
Overview
zerogpu is a plugin marketplace for the Build phase that routes classification, summarization, extraction, and short chat to ZeroGPU small/nano models via 14 auto-invoked skills.
What is this marketplace?
- 14 auto-invoked skills for classification, summarization, extraction, and short chat
- Routes NLP workloads to ZeroGPU small and nano models for cost optimization
- Tagged flows: PII, NER, JSON extraction, follow-ups, and summarization
- Productivity-category plugin sourced from agents/claude
- Designed for automatic invocation rather than manual skill picking each turn
- 1 plugin (zerogpu-router)
- 14 auto-invoked skills stated in marketplace description
- Plugin tags include nlp, classification, summarization, pii, ner, cost-optimization
Community signal: 77 GitHub stars.
What problem does it solve?
Solo builders overpay and slow down agents by using large models for every label, summary, and PII scan.
Who is it for?
Agent founders optimizing inference cost on classification, summarization, NER, PII redaction, and lightweight chat sidecars.
Skip if: Projects with no external API allowance or teams that need exclusively on-device models with zero cloud routing.
What do I get? / Deliverables
After registration, repetitive NLP side tasks can auto-route to ZeroGPU small models while your main agent keeps frontier models for hard reasoning.
- 14 routed NLP skill hooks (classification, summarization, extraction, chat)
- Cost-optimized path for PII, NER, and JSON extraction side tasks
- Auto-invocation behavior for tagged productivity workflows
Plugins in this marketplace
1 plugin — install individually after you add the marketplace.
Recommended Marketplaces
Journey fit
Routing inference to external small models is an integration decision made while wiring the agent product, before you optimize bills in production. zerogpu-router is a model-routing layer with fourteen auto-invoked skills—classic LLM vendor integration, not frontend polish.
How it compares
LLM routing integration marketplace, not a self-hosted weights repo or a design-system skill.
Common Questions / FAQ
Who is Zerogpu for?
Solo builders shipping Claude Code agents who want automatic routing of routine NLP tasks to ZeroGPU small and nano models.
When should I use Zerogpu?
Use it when you integrate cost-sensitive classification, summarization, entity or PII extraction, or short replies alongside your primary coding agent.
How do I add Zerogpu to my agent?
Add the Zerogpu Claude marketplace, enable the Zerogpu-router plugin from agents/claude, and configure ZeroGPU credentials per their docs.