Together AI integration

What it does

The Together AI integration lets your agency call open-source LLMs hosted on Together's serverless inference platform from inside any workflow. Connect a client's Together AI API key once and your workflows can run chat completions against Llama 3, Mixtral, Qwen, and the full Together model catalog, generate text completions for prompt-style models, produce embeddings for retrieval pipelines, and surface the model catalog so end users can pick the right model and the right cost ceiling for the job.

Connect a Together AI account

1. Open your workspace in TaskJuice and navigate to Connections.
2. Choose Together AI and click Connect.
3. In a new tab, open the Together AI API keys page signed in as the client (or as your agency, if the client has delegated key creation to you).
4. Click Create new key, give it a name that identifies the workspace, and copy the value.
5. Paste the key into TaskJuice and save the connection.

To rotate or revoke the key later, return to the Together AI API keys page, delete the entry, create a new key, and update the TaskJuice connection.

Triggers

Together AI does not publish a public webhook surface, so the integration is action-only per the AI-provider exception in the catalog policy. Use a Schedule, a polling trigger from another app, or an inbound webhook from a different source to invoke Together AI actions inside a workflow.

Actions

• togetherai/create-chat-completion sends a message-formatted prompt to a chosen Together AI chat model with optional temperature, max-tokens, top-p, response-format, and stop-sequence controls.
• togetherai/create-completion sends a raw text prompt to a completion-style model and returns the generated text along with usage counts.
• togetherai/create-embedding returns an embedding vector for an input string using a chosen Together AI embedding model, ready to be written to a vector store.
• togetherai/list-models returns the model catalog available to the connected key, useful for surfacing a model picker or branching on model availability.

Known limitations

• The integration calls the Together AI REST API with the connected key. Per-key rate limits, monthly spend caps, and model-access entitlements are governed by the client's Together AI account, not by TaskJuice. When Together AI returns a 429 or a 5xx response, TaskJuice surfaces it as a retryable rate-limit or provider error and applies backoff.
• Streaming responses are not exposed. Each Create Chat Completion and Create Completion call runs as a single request and resolves only after the full response is returned.
• Tool-use, vision, and JSON-mode parameters beyond the documented response_format field are not first-class fields on the chat-completion action yet. The messages array accepts the full content schema, so structured content blocks can still be passed through directly.
• The integration is action-only. To react to events in real time, use a Schedule trigger, a polling trigger from another app, or an inbound webhook from a different source and call Together AI inside the workflow.