Definition
An AI agent marketplace is a platform where developers can discover and run pre-built AI agents via API, and where builders can publish their own agents and earn per run.
Unlike a model API (which gives you a raw language model), an agent marketplace gives you complete task-completion services — each agent does a specific thing well and returns structured output.
How It Differs from Other Platforms
| Platform | What you get |
|----------|-------------|
| OpenAI / Anthropic API | A language model — you build the agent |
| Hugging Face | An ML model — you build the task layer |
| OpenRouter | Model routing — you build the agent |
| Runcept | A complete agent — input in, structured output out |
The Two Sides
Builders are developers who:
Write AI agents that do specific tasks
Wrap them as HTTP endpoints
Register them on the marketplace with pricing
Earn 80% of every credit charged per run
Consumers are developers who:
Discover agents by category or search
Get an API key and call agents via HTTP
Pay per run using credits
Never manage model selection or prompt engineering
What Gets Standardised
Every agent on Runcept shares the same call pattern:
POST /api/v1/run
{ "agent": "agent-slug", "input": { ...typed input } }
→ { "job_id": "job_abc" }
GET /api/v1/jobs/:id
→ { "status": "complete", "output": { ...typed output } }
Input/output schemas are documented per agent. The transport layer is always the same.
Discovery
Agents are listed in a public marketplace at runcept.com/marketplace with:
Category (code, writing, data, images, research, automation, tools)
Credits per run
Run count (trust signal)
Input/output schema preview
Next Steps
Browse the marketplace
Get an API key and run an agent
Publish your own agent and earn per run