Fireworks AI
for LuumenAI
Check Fireworks models and deployments, then run inference as an approved step
Connect Fireworks AI and Luumen can read your account's models and deployments on request — which deployment IDs exist, how a deployment is configured, whether a model resource is ready. Inference calls such as chat completions, embeddings, and reranking are prepared as a step you review before it runs. Authentication is by API key.
The Fireworks AI toolbox
9 tools: 5 read, 4 write. Reads answer instantly. Writes require approval by default. Everything is logged.
- ReadGet DeploymentReturn configuration and serving status for one Fireworks deployment without scaling or modifying it.
- ReadGet ModelReturn metadata and status for one model resource in a Fireworks account; this does not test whether the model can serve inference.
- ReadList AccountsReturn one page of Fireworks accounts accessible to the API key so an agent can obtain account IDs for model and deployment discovery.
- ReadList DeploymentsReturn one page of deployment resources in a Fireworks account for deployment ID discovery and serving-capacity and status inspection.
- ReadList ModelsReturn one page of model resources in a Fireworks account for model-ID discovery and status inspection; availability for inference is not guaranteed.
- WriteCreate Chat CompletionGenerate one non-streaming OpenAI-compatible chat completion from role-based messages using a caller-supplied Fireworks model or deployment ID. Approval by default
- WriteCreate EmbeddingsCreate embedding vectors for one or more caller-provided inputs using a caller-supplied Fireworks embedding model. Approval by default
- WriteGenerate Model ResponseGenerate one non-streaming Responses API result, optionally continuing a stored response or using function, MCP, or SSE tools; storage is off by default. Approval by default
- WriteRerank DocumentsRank a provided document list by relevance to a query using an optional caller-supplied Fireworks reranker model. Approval by default
One prompt, start to finish
What a governed Fireworks AI run looks like inside Luumen.
Questions
Authorize once with API token. Luumen lists the scopes each action needs before you approve the connection, and credentials never appear in the chat.
Read actions answer immediately. Anything that writes — create chat completion, create embeddings, generate model response, rerank documents — is shown as a plan and requires approval by default. Administrators configure that per tool, so you decide exactly which actions can ever run unattended.
You decide. Actions are granted per agent, skill, and team, and per environment — production is not staging. Read access can be broad while writes stay narrow.
Every call to Fireworks AI — read or write, approved or declined — is recorded with the actor, the input, and the result, and can be linked to the ticket or change record.
Often connected alongside
All integrations
Adobe Marketing Agent MCPAdobe Experience Platform context, sandbox switches, and async task status
AffindaRead parsed documents, collections, and annotations — change them with approval
Agent MailGive the agent its own mailbox for reading threads and sending updates
AI/ML APIInspect assistants, threads, and runs — and change them as approved stepsPut Fireworks AI to work with Luumen
Connect in minutes. Every action scoped, approved, and audited from day one.