Connect your model client. Inspect your first request.
Configure your provider, send a small request through the model gateway, and inspect its response and usage. Connecting an agent to another agent or an MCP tool has its own quickstart below.
This gateway example uses HTTP and needs no Praesidia SDK. Use your own provider account and key; provider calls incur provider charges. No credit card required.
curl --fail-with-body "${PRAESIDIA_GATEWAY_ORIGIN}/openai/v1/chat/completions" \
-H "Authorization: Bearer ${PRAESIDIA_APPLICATION_KEY}" \
-H 'Content-Type: application/json' \
--data "{\"model\":\"${PRAESIDIA_MODEL}\",\"messages\":[{\"role\":\"user\",\"content\":\"Say hello in one sentence.\"}]}"From the model gateway quickstart. Set the three variables to your gateway origin, your application key and a model enabled for your route; no SDK is needed.
From provider setup to a request you can inspect
Follow these steps for the command above. Start with a small request and a model enabled on your provider account.
- 1
Configure the provider. Add your provider key in Configure → Integrations → LLM Configurations, then create an application and its gateway application key. Provider and key setup
- 2
Send the request. Get your workspace's gateway address from your administrator, set the three variables, and run the command above. Supported request examples
- 3
Inspect the response and usage. Check the model response and find the matching application and organization in your workspace's usage and audit activity. What to check
Once the request succeeds, review budget scopes and enforcement and test the controls available to your plan before increasing traffic. Troubleshoot the first request.
What you can use today, and how to get the rest
- Agent and tool connections
- To govern agent-to-agent or agent-to-MCP calls, follow the connection quickstart. It covers registering the participants, creating a connection and testing a policy. Your deployment operator supplies the runtime credential. Then use the action-to-evidence guide to inspect a protected call.
- Frameworks and runtimes
- LangGraph, OpenAI Agents SDK, Google ADK and more, each with the recipe we have validated and what it covers: frameworks. The CrewAI release is in validation. For coding agents, workflow tools and MCP clients, see the runtime catalog.
- SDKs
- The TypeScript SDK source is public. TypeScript and Python packages are not yet published to npm or PyPI; design partners receive release artifacts directly. Runtime installation covers installing a provided release. Ask for access.
- The offline verifier
- Apache-2.0 licensed; the source repository is private. Ask your Praesidia contact for verifier access or a packaged build; see how verification works. No public registry release is assumed.
- Plans
- The signed audit trail, rule-based guardrails, API keys and bring-your-own-key model configuration are in every plan, including the free one. Protected-action evidence reads are on Advanced or Enterprise where enabled; registered protected HTTP execution also requires Enterprise approval workflows. Approving a held call is in every plan, and multi-step approval workflows are Enterprise. Compare plans.
Read next: choose your first outcome · all docs · the evidence · runtime security
Start with one small request
Create a workspace, configure your provider and follow the model gateway quickstart.