GSE AI Server API

This package provides an API for connecting to UB GSE's AI server (powered by LocalAI). See the Computational Literacies Lab handbook page for details about the server.

The documentation provides details on using the API.

Quickstart

from gseai import GSEAIServer

with GSEAIServer("your-api-token") as server:
    # List available models
    models = server.list_models()

    # Simple single-turn chat
    response = server.chat("model-id", "What is machine learning?")

    # Multi-turn chat using the OpenAI-compatible API
    response = server.chat_completions(
        "model-id",
        messages=[{"role": "user", "content": "What is machine learning?"}],
    )

Job queue

Large models can take longer than the server's HTTP timeout to respond. The job queue lets you submit a request and collect the result later — no need to keep a connection open.

from gseai import GSEAIServer

with GSEAIServer("your-api-token") as server:
    # Submit a job and get an ID immediately
    job = server.submit_job("my-job", "qwen2.5-coder", "Explain transformers.")
    job_id = job["job_id"]

    # Come back later and wait for it
    result = server.wait_for_job(job_id)
    print(result["result"])

    # File-in jobs (transcribe, image_edit, etc.)
    job = server.submit_file_job("transcribe-lecture", "transcribe", "whisper-1", "lecture.mp3")

    # Binary-out jobs (speech, images) — fetch the file when done
    audio = server.get_job_result(job_id)
    open("output.mp3", "wb").write(audio)

The CLI supports the same workflow:

# Submit and walk away
gseai queue submit qwen2.5-coder "Explain transformers." -n my-job
# a3f1c7d2-...

# Check status, then fetch result when done
gseai queue status a3f1c7d2-...
gseai queue wait a3f1c7d2-...

# Or submit and wait in one step
gseai queue run qwen2.5-coder "Explain transformers." -n my-job

# File-in jobs
gseai queue upload whisper-1 lecture.mp3 --job-type transcribe -n lecture
gseai queue wait <job-id>