Chat

gumloop chat completions create is the terminal counterpart of the Python SDK call client.chat.completions.create(...). Every flag maps 1:1 to the matching SDK kwarg.

Create a completion

gumloop chat completions create "Capital of Canada?" -m claude-sonnet-4-5

By default, output streams to your terminal when stdout is a TTY. When it isn’t, for example when piped to a file or another command, the CLI buffers the output into a single response. Pass --stream or --no-stream to be explicit. A structured request can also set stream; explicit flags take precedence.

Flag	Description
`-m`, `--model`	Required. Model slug (for example `claude-sonnet-4-5`, `gpt-4o-mini`, `gemini-2.5-pro`). Chat needs a concrete model, so `gumloop-chew` (the Auto router) isn’t valid here. Preview a choice with `POST /models/route` first, then complete.
`-s`, `--system`	System message prepended to the conversation. Repeatable.
`--message-stdin -`	Read the user message from stdin instead of the positional argument.
`--max-completion-tokens`	Cap on completion tokens.
`--temperature`	Sampling temperature.
`--modality`	Output modality. Repeatable, for example `--modality image --modality text` for image-generation models.
`--schema-file`	Path to a JSON Schema file. Sent as `response_format={"type": "json_schema", ...}` for structured output.
`--schema-name`	Name to attach to the schema (defaults to `schema`).
`--args-file`, `--args-json`, `--args -`	Full SDK request as a JSON object, from a file, inline JSON, or stdin. Choose one source.
`--stream` / `--no-stream`	Force or suppress streaming. When neither a flag nor the request JSON specifies streaming, streams only if stdout is a TTY and `--json` is absent. Explicit flags override the request JSON.
`--json`	Print the response as JSON. With `--stream`, emits newline-delimited JSON (one chunk per line).

Pipe from stdin

cat draft.md | gumloop chat completions create -m claude-sonnet-4-5 \
  --message-stdin - \
  --system "Summarize the input in three bullets."

Structured output

gumloop chat completions create "Return JSON with the capital of Canada." \
  -m claude-sonnet-4-5 \
  --schema-file ./capital.schema.json \
  --schema-name capital \
  --json

capital.schema.json:

{
  "type": "object",
  "properties": { "capital": { "type": "string" } },
  "required": ["capital"],
  "additionalProperties": false
}

Image generation

gumloop chat completions create "A red maple leaf on white" \
  -m gpt-image-2.5-flare \
  --modality image --modality text \
  --json

The response carries one or more image attachments on choices[0].message.images. Each entry is a data URL the caller can decode or render directly. Human output prints text. Use --json to retain images, including choices[0].delta.images in streamed chunks. Image generation options belong in image_config in a structured request.

Structured requests and vision

--args-file accepts the SDK’s full request shape: conversation history, multimodal message content, tool definitions and tool results, reasoning options, and provider-supported generation settings. Explicit command flags override corresponding request fields. Check the selected model’s supported inputs and outputs.

gumloop chat completions create --args-file request.json --json

{
  "model": "VISION_MODEL_ID",
  "messages": [{
    "role": "user",
    "content": [
      {"type": "text", "text": "Describe this image"},
      {"type": "image_url", "image_url": {"url": "data:image/png;base64,REPLACE_WITH_IMAGE_BYTES"}}
    ]
  }]
}

Replace the placeholders with a vision-capable model ID and an encoded image. Vision analyzes an input image. Text-to-image generation uses an image-capable model and the image output modalities described above. Reference-image editing additionally requires the backend image-input adapter and a supported image-generation model. Supply inline base64 image_url content with image output modalities. Remote image URLs and storage paths are refused by that adapter; MIME, decoded size, and model limits apply. This capability depends on backend deployment rather than the CLI installation alone.

Streaming with machine output

--stream --json emits ndjson, one full chat.completion.chunk per line, so consumers can stitch deltas without re-parsing the full SSE wire format.

gumloop chat completions create "stream me" -m claude-sonnet-4-5 --stream --json

Without an explicit streaming setting, the CLI streams to TTYs and uses unary output for redirects and pipes.

SDKs

Gumloop for terminal

Create a completion

Pipe from stdin

Structured output

Image generation

Structured requests and vision

Streaming with machine output

​Create a completion

​Pipe from stdin

​Structured output

​Image generation

​Structured requests and vision

​Streaming with machine output

Create a completion

Pipe from stdin

Structured output

Image generation

Structured requests and vision

Streaming with machine output