Gemini and Dome
Run Gemini on Vertex AI, with Pro kept for the work that needs it. Dome calls Gemini through Vertex AI with your service account. Pool Gemini 3.8 Flash with an older Flash behind it, and agents keep working when one model is unavailable. Rules keep Gemini 3.1 Pro to the people who need it.
Dome provides model brokeringand routing for Gemini.
A fallback model behind Flash
Pool Gemini 3.8 Flash with Gemini 2.5 Flash behind it. When the first fails, the call goes to the second.
Pro for the teams that need it
Tag the Pro connection with its tier. A rule refuses it unless the person the agent acts for is in ml-research.
A ceiling on spend
Cap Pro at $400 of provider spend a month. Once it's spent, Pro calls are refused until the window resets.
Gemini on Vertex AIin five steps.
Connect each model with your Google Cloud project and service account, pool the Flash models, and point your agents at the Gateway.
Connect Gemini 3.8 Flash
Pass the service account's JSON key as the credential. Dome stores it in its vault and exchanges it for Google access tokens at call time.
shelldome models add gemini-38-flash \ --provider google \ --model gemini-3.8-flash \ --provider-config '{"project":"acme-agents","location":"us-central1"}' \ --api-key "$(cat vertex-sa.json)" \ --gateway prod-gatewayConnect the fallback and Pro
One connection per model. Tag Pro so a rule can name it.
shelldome models add gemini-25-flash --provider google --model gemini-2.5-flash \ --provider-config '{"project":"acme-agents","location":"us-central1"}' \ --api-key "$(cat vertex-sa.json)" --gateway prod-gateway dome models add gemini-pro --provider google --model gemini-3.1-pro \ --provider-config '{"project":"acme-agents","location":"us-central1"}' \ --api-key "$(cat vertex-sa.json)" --attributes '{"tier":"pro"}' \ --gateway prod-gatewayPool the Flash models
Members are tried in priority order, and agents send the pool's name.
shelldome models pool create gemini-flash \ --failover-max all --gateway prod-gateway dome models pool member add gemini-flash gemini-38-flash --priority 0dome models pool member add gemini-flash gemini-25-flash --priority 1Point your agent at the Gateway
Any OpenAI-compatible client works. Dome translates the request to Gemini's format.
pythonfrom openai import OpenAI client = OpenAI(base_url=f"{GATEWAY_URL}/v1", api_key=DOME_AGENT_KEY)client.chat.completions.create( model="gemini-flash", messages=[{"role": "user", "content": "Extract the invoice fields."}],)Cap the spend
A model quota counts every call to that connection, through pools or directly.
shelldome quotas set --subject model --model gemini-pro \ --unit provider_usd --limit 400 --window monthly
Gemini Pro for research,refused for everyone else.
Any connection tagged pro is refused unless the person the agent acts for is in ml-research. The group arrives in their identity token.
cedar
forbid (principal, action == Dome::Action::"llm:invoke", resource is Dome::LLMModel)when { resource has tier && resource.tier == "pro" }unless { principal has act_as && principal.act_as.groups.contains("ml-research")};One call,two outcomes.
Switch the caller or the argument and watch the same call decide differently, with every decision landing in audit.
Acting for
agent research-agent · acting as p.nakamura
llm:invoke(model: "gemini-pro")
- Agentresearch-agent is registered and active
- Callerp.nakamura verified, groups: ml-research
- RulePro is open to ml-research
- Quota$150 of $400 this month
DecisionAllowed
Bringing it alltogether.
Connecting Gemini 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
GeminiModel
Google GeminiProviderCommon questionsabout Gemini.
More of whatDome works with.
Provider
Anthropic
The Claude API behind the Model Broker: the key held in Dome, every call authorized, metered and audited.
MCP server
GitHub
The GitHub MCP server behind the Tool Gateway, with rules that decide each call on its owner and repository.
Identity
Okta
Okta tokens verified on every agent call, so rules and audit name the person each agent acted for.
Runtime
LangGraph
LangGraph agents with their model calls 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.
AI models for enterprise agents: failover, rules and quotas
Every model your agents call goes through the Model Broker. Pool providers for failover, decide who may call each model, and cap what it costs.