Platform tour
Everything you need to ship AI. Nothing you have to maintain.
Agents, knowledge, tools and tracing sit behind one endpoint. Configure them in a dashboard, call them from any language, and read back exactly what happened on every request.
14 days free · No credit card · Bring your own model keys
01 · Developer API
One endpoint for every agent
Authenticate with your key and POST to /v1/chat/stream from any language that speaks HTTP. There is no SDK to install, and the response carries the reply plus the model, tokens, latency and cost it took to produce it.
How the API works02 · Tools & data
Agents that reach into your stack
Point an agent at an MCP server and it discovers what is there — databases, calendars, CRMs, internal APIs — then calls the right tool when a question needs one. Add an endpoint upstream and it is available on the next request.
See MCP in action03 · Tracing
See exactly what the agent did
Every call becomes an OpenTelemetry trace: auth, retrieval, tool calls, the model, the response — with latency, tokens and cost on each span. Debug a bad answer in the dashboard, or export the spans to the backend you already run.
Look inside a trace04 · Knowledge / RAG
Answers grounded in your own material
Upload PDFs, write FAQs, crawl your site. Retrieval runs on every call, so agents answer from your documents instead of the open internet — with no retraining and no ingestion pipeline to keep alive.
How retrieval works05 · Test before you ship
Try it before your users do
Chat with any agent from the dashboard, watch which tools it reaches for, adjust the prompt or the model, and hand it a production key only once it behaves the way you want.
Test an agent firstAlso included
Three more things you won't have to build.
A chat widget for your site
Turn any agent into a live chatbot on WordPress, Shopify or your own front end, styled to match the rest of the page.
See the widgetAI Workspace for your team
A shared chat space grounded in your own documents, so the team asks the same agents your product already runs on.
See the workspaceYour cloud, when it matters
Run SentientOne inside your own AWS, Azure, GCP or on-prem estate. Prompts, documents and traces never leave your boundary.
Read about self-hostingSee it running on your own data.
Start the trial, wire one agent into a real screen, and judge the whole platform with something working in front of you.
14 days free · No credit card · Cancel anytime