SynapsAI Cloud
SynapsAI Cloud is the workspace for running AI in production: models, agents, and knowledge bases, with an OpenAI-compatible API you can drop into an existing stack. You can launch most kinds of AI model, not only chat LLMs: embeddings, images, audio, video, classification, question answering, rerank, and the rest of the Hugging Face pipeline set we list under Supported tasks. Point at a Hub repo (or an artifact), launch, and call a private endpoint. The platform is in public beta. Share ideas on feedback and feature requests.Quickstart
Create an account, launch a model, and send the first API call.
Plug into your stack
LangChain, LlamaIndex, OpenAI SDKs, Gradio, and FastAPI.
The platform
Everything in the dashboard is one account: launch and scale models, try them in the playground, attach knowledge bases to agents, then call the same deployments from your code.Models
Launch almost any supported pipeline from Hugging Face, an artifact, or a custom container. Scale workers and call a private endpoint.
Agents
Saved prompts, tools, MCP servers, and knowledge bases on top of a deployed model.
Knowledge bases
Upload or sync documents, then search them from agents, MCP, or the vector store API.
Playground
Try a model or agent in the dashboard before you wire up production clients.
How a workload runs
1
Launch a model
Point at a Hugging Face repo, an artifact, or a custom image. Set workers, timeout, and pricing in the wizard.
2
Call the API
Use
synapsai-python, cURL, or any OpenAI-compatible client against https://api.synapsai.cloud/v1 with your API key. The model field is your deployment id.3
Add retrieval and agents
Index documents in a knowledge base and optionally wrap the model in an agent.
Where to go next
Core concepts
Lifecycle, workers, credits, and naming (knowledge base vs vector store).
Examples
Copy-paste requests for chat, embeddings, images, audio, and more.
Migrate from OpenAI
Change the base URL and API key; keep your existing client.
API reference
Authentication, errors, and endpoint groups.
Account
Keys, billing, usage, quotas, and teams.
Supported tasks
The full list of pipelines you can launch — chat is only one of them.
Need help? See Troubleshooting, the status page, or contact support.