Announcing Dome Managed Agents.

Try it now

SambaNova API and Dome

Open models on SambaNova, with every call governed. SambaNova labels some of its models preview. Connect it to Dome, tag those connections stage: preview, and production agents can't reach them while evaluation agents can. Fireworks AI covers DeepSeek V3.1 when SambaNova is down.

Dome provides model brokeringand routing for SambaNova.

Agents never hold the key

SambaNova's key lives with Dome. Agents call the Gateway with their own credentials, and Dome adds SambaNova's on the way out.

Preview models stay in testing

SambaNova marks some models as preview. Tag those connections and a rule keeps production agents off them.

DeepSeek V3.1, twice over

Fireworks AI serves DeepSeek V3.1 as well. Pool the two and calls SambaNova can't answer go to Fireworks.

Connect SambaNovain three steps.

Add a SambaNova connection, give it a Fireworks AI fallback, and route agents through Dome.
  1. Add the connection

    The provider id is sambanova. Dome uses https://api.sambanova.ai/v1 and SambaNova's own model ids.

    shell
    dome models add deepseek-sambanova \  --provider sambanova \  --model DeepSeek-V3.1 \  --api-key "$SAMBANOVA_API_KEY" \  --attributes '{"stage":"production"}' \  --gateway prod-gateway
  2. Pool it with Fireworks AI

    Add accounts/fireworks/models/deepseek-v3p1 as a second connection, ranked after SambaNova.

    shell
    dome models pool create deepseek-v3.1 \  --failover-max all --gateway prod-gateway
    dome models pool member add deepseek-v3.1 deepseek-sambanova --priority 0dome models pool member add deepseek-v3.1 deepseek-fireworks --priority 1
  3. Point your agent at the Gateway

    OpenAI-compatible code needs a new base URL and the agent's Dome key, and nothing else changes.

    python
    from openai import OpenAI
    client = OpenAI(base_url=f"{GATEWAY_URL}/v1", api_key=DOME_AGENT_KEY)client.chat.completions.create(    model="deepseek-v3.1",    messages=[{"role": "user", "content": "Answer this customer's billing question."}],)

No preview modelsin production.

Applied to a production agent, this refuses any connection tagged stage: preview. Evaluation agents without the rule can still try them.
cedar
forbid (principal, action == Dome::Action::"llm:invoke", resource is Dome::LLMModel)when { resource has stage && resource.stage == "preview" };

Bringing it alltogether.

Connecting SambaNova to registered agents, tools, and identity in Dome completes a governed agent application.
Dome

Acting for

Control point

GatewayEvery call decided and audited
RulesGuardsQuotas

Common questionsabout SambaNova.

How does Dome connect to SambaNova?
Through SambaNova's OpenAI-compatible API at https://api.sambanova.ai/v1. Dome holds the key and presents it as a Bearer header.
Which SambaNova models can I use?
Any id SambaNova lists, such as DeepSeek-V3.1, Meta-Llama-3.3-70B-Instruct or gpt-oss-120b, written exactly as SambaNova spells it.
How do I keep preview models out of production?
Tag preview connections with an attribute such as stage: preview. A rule on each production agent refuses them.

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

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.

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.

Every agent you run, under one system of control.