Together
Run Together models with streaming, image input, reasoning, structured output, web search, and retry configuration.
Together retires serverless model IDs on a rolling schedule. The linked cookbook examples may reference retired IDs. Before running an example, choose a current model from the serverless catalog and verify the capabilities required by that example.
| Example | Description |
|---|---|
| Together Basic | Run a current Together chat model with sync, streaming, and async calls. |
| Together Image Agent | Describe an image from a URL with a current Together vision model. |
| Together Image Agent Bytes | Send local image bytes to a current Together vision model and stream the description. |
| Together Image Agent With Memory | Keep image context across turns with a current Together vision model and an in-memory database. |
| Retry | Review retry settings and why invalid model IDs cannot reliably exercise the retry path. |
| Together Structured Output | Request JSON through Together and validate a MovieScript with Pydantic. |
| Tool Use | Give a function-calling Together model web search tools and run it sync, streaming, and async. |
| Reasoning alternatives | Current alternatives for the retired Together Reasoning Agent example. |