akka optimize training endpoints send
Send one input to a model at an endpoint
Synopsis
Send one input to a model and print its response.
ENDPOINT takes an ID or an unambiguous prefix of it. The request goes through the training service, which authenticates it, so no separate credentials are needed.
--input takes the text to send, and --file reads it from a file instead, or from standard input where the path is "-".
A model trained here is asked under the instructions it was trained with, which "endpoints get" shows. --instructions replaces them for this request.
This sends one input and keeps no history. To drive the model from an application or a test script, use "endpoints proxy", which opens a local address the endpoint answers at in the OpenAI format.
If the endpoint serves one model, that model is used. If it serves several, choose one with --model. The value may be the displayed model name (including a shortened adapter) or the model’s registry reference, NS/NAME:vN, as "endpoints get" shows it.
--tools offers the model the tools in a file, an OpenAI "tools" array. An endpoint returns tool calls only for a model whose architecture has a tool-call parser, which "endpoints info" shows, and refuses a request with tools otherwise.
The answer is printed as it is produced. Reasoning the endpoint returns apart from the answer is printed to standard error, and each tool call the model asks for is printed after the answer. Under -o json the model, the output, the reasoning, the tool calls and the token usage are returned instead.
akka optimize training endpoints send ENDPOINT --input TEXT [flags]
Examples
# Send one input
akka optimize training endpoints send 970076e8 --input "Payment failed twice, card declined."
# Ask under instructions of your own, and take the same answer every time
akka optimize training endpoints send 970076e8 --input "..." \
--instructions "Reply with one label." --temperature 0
# Read the input from a file, against one of several models
akka optimize training endpoints send 970076e8 --file ticket.txt --model Qwen/Qwen3-8B:run-7
# Offer the model tools
akka optimize training endpoints send 970076e8 --input "Weather in Paris?" --tools tools.json
Options
--file string Read the input from a file, or from standard input where the path is "-"
-h, --help help for send
--input string Text to send to the model
--instructions string Instructions to send in place of the ones the model was trained with
--max-tokens int Maximum tokens in the answer; zero leaves it to the deployment
--model string Model to use by name or registry reference when the endpoint serves several
--temperature float Sampling temperature; unset leaves it to the deployment
--tools string File holding an OpenAI "tools" array to offer the model
Options inherited from parent commands
--cache-file string location of cache file (default "~/.akka/cache.yaml")
--config string location of config file (default "~/.akka/config.yaml")
--context string configuration context to use
--disable-prompt Disable all interactive prompts when running akka commands. If input is required, defaults will be used, or an error will be raised.
This is equivalent to setting the environment variable AKKA_DISABLE_PROMPTS to true.
--force Skip the confirmation prompt
--jq string Filter the JSON output with a jq expression
--json string Output JSON with only these fields; --json= lists the fields a command offers
--no-color Print without color, as NO_COLOR does
-o, --output string set output format to one of [text,json,json-compact,go-template=] (default "text")
--page-mode string the mode for paging, either paged, buffered or auto. (default "auto")
-q, --quiet set quiet output (helpful when used as part of a script)
--region string Region to use if the project has more than one region
--timeout duration client command timeout (default 10s)
--trainer-url string Trainer service URL (default TRAINER_URL, then AKKA_OPTIMIZE_URL, then http://localhost:9001)
--url string Optimize service URL (default OPTIMIZE_URL, then AKKA_OPTIMIZE_URL, then http://localhost:9010)
--use-grpc-web use grpc-web when talking to Akka APIs. This is useful when behind corporate firewalls that decrypt traffic but don't support HTTP/2.
--verbose set verbose output
SEE ALSO
-
akka optimize training endpoints - Serve a base model or a candidate to try it by hand