Claude API and Dome
One Anthropic key serves every agent, and every call is governed. Connect the Claude API to Dome once. Agents reach Claude with Dome keys of their own, Anthropic's key stays in Dome's vault, and rules decide each call before Anthropic sees it.
Dome provides model brokeringand routing for Anthropic.
Anthropic's key, out of reach
Dome injects the x-api-key header at call time. Every read of the key is audited, and no agent ever holds it.
Agents authenticate as themselves
Each agent calls with its own Dome key. Revoke the agent and its calls stop, with the Anthropic key untouched.
A ceiling on every loop
Calls are metered to the agent and the person it acted for. A quota halts a runaway Claude loop at the limit you set.
Connect Anthropicin two steps.
Add the connection, then change your agent's base URL. Existing Anthropic and OpenAI SDK code keeps working.
Add the connection
Any Claude model id works. Add one connection per model you want agents to reach.
shelldome models add claude-fable \ --provider anthropic \ --model claude-fable-5-1 \ --api-key "$ANTHROPIC_API_KEY" \ --gateway prod-gatewayPoint your agent at the Gateway
The Anthropic SDK takes the Gateway URL as its base. OpenAI-compatible clients use the same URL with /v1.
pythonimport anthropic client = anthropic.Anthropic(base_url=GATEWAY_URL, api_key=DOME_AGENT_KEY)client.messages.create( model="claude-fable", max_tokens=1024, messages=[{"role": "user", "content": "Draft a reply to this ticket."}],)
Frontier modelsby group.
Tag a connection with an attribute and write a rule against it. This one keeps frontier connections to one team, whichever agent is calling.
cedar
forbid (principal, action == Dome::Action::"llm:invoke", resource is Dome::LLMModel)when { resource has tier && resource.tier == "frontier" }unless { principal has act_as && principal.act_as.groups.contains("ml-research")};Bringing it alltogether.
Connecting Anthropic to registered agents, tools, and identity in Dome completes a governed agent application.
Acting for
Agent
Control point
GatewayEvery call decided and audited
RulesGuardsQuotas
Models
AnthropicProvider
Claude OpusModelCommon questionsabout Anthropic.
More of whatDome works with.
Model
Claude Fable
Fable 5.1 from Anthropic and Amazon Bedrock in one failover pool, open to one group and capped by quota.
MCP server
GitHub
The GitHub MCP server behind the Tool Gateway, with rules that decide each call on its owner and repository.
Identity
Microsoft Entra ID
Entra access tokens verified on every agent call, so rules read app roles and audit names the person.
Runtime
OpenAI Agents SDK
OpenAI Agents SDK agents with their models on the Model Broker and their MCP tools on the Tool Gateway.
Client
Claude Code
Claude Code on a Dome Gateway with per-developer sign-in, rules on every tool call, and audit by name.
Agent service
TinyFish
TinyFish's web agents behind the Tool Gateway, with rules that decide each run on the site it targets.
Talk with ourFDE team.
Or start free, and register your first agent in minutes. Connect it through a gateway, give it rules, and see every call it makes in one audit trail.
Talk with our FDE team
Our engineers design, build and run governed agents for the part of your business you want to change. You see what they deliver.
LLM providers for AI agents: one governed path to every model
Connect a provider once. Its key stays in Dome, and every agent call to it is authorized, metered and audited.