Skip to main content
Create a workspace, upload a dataset, fine-tune a foundation model in FineTune Studio, deploy it, and call the LLMTune API.
1

Create your account and workspace

  1. Go to https://llmtune.io/login.
  2. Sign up or log in with your preferred method.
  3. The first time you sign in, a default workspace is created for you. You can create additional workspaces later from the dashboard.
2

Create an API key

  1. In the dashboard, open API Keys.
  2. Click Create API Key.
  3. Give the key a descriptive name (for example: production-backend or staging-testing).
  4. Copy the key and store it securely – you won’t be able to see it again.
3

Upload and prepare data (Dataset Hub)

  1. Navigate to Dataset Hub from the main navigation.
  2. Choose Upload dataset or Connect source.
  3. Supported options today include:
    • Uploading files (JSONL, CSV, TXT)
    • Using preconfigured playground / demo datasets
  4. LLMTune validates the format and shows a preview.
  5. (Optional) Apply tags, notes, and quality checks. PII detection and cleaning are handled automatically where enabled.
4

Launch a fine-tune (FineTune Studio)

  1. Open FineTune Studio from the product navigation.
  2. Pick a base model from the catalog. Cards highlight provider, context length, and recommended use cases.
  3. Select a dataset (or blend multiple datasets) from Dataset Hub.
  4. Choose a training method:
    • SFT – Supervised Fine-Tuning
    • DPO – Direct Preference Optimization
    • PPO – Policy optimization with rewards
    • RLAIF – RL with AI feedback
    • CTO – Controlled Tuning Optimization
  5. Configure key hyperparameters:
    • Learning rate
    • Batch size
    • Epochs
    • Evaluation cadence
  6. Choose your compute model:
    • Traditional (single instance or GPU cluster)
    • Federated (distributed compute)
  7. Confirm the cost estimate and click Launch training.
  8. Watch live telemetry: tokens/sec, loss curves, queue position, and status.
5

Promote to an endpoint (Deploy)

  1. When a run completes, open Deploy (Deploy Control) from the product navigation or from the training run details.
  2. Click Promote to endpoint.
  3. Choose environment (staging or production).
  4. Optionally configure rollout strategy and traffic splitting.
  5. Save the deployment – LLMTune links it to the original dataset, training job, and workspace.
6

Call the LLMTune API

  1. Once you have a deployed model ID, you can call the public LLMTune API.
  2. Base URL for the public API is:
https://api.llmtune.io/v1
  1. Run your first inference call:
curl https://api.llmtune.io/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/Llama-3.3-70B-Instruct",
    "messages": [
      { "role": "system", "content": "You are a precise assistant." },
      { "role": "user", "content": "Summarize this support ticket." }
    ],
    "temperature": 0.7,
    "max_tokens": 400
  }'
1

Evaluate your model (Evaluate Suite)

  1. From the training job or deployment, click Evaluate.
  2. Use the evaluation interface to:
    • Run single prompt checks
    • Compare with base model side-by-side
    • Run batch evaluations over many prompts
    • Inspect the results dashboard for quality trends
  3. Use evaluation results to decide whether to promote, adjust training data, or launch a new run.
Continue with the Fine-Tuning Guide for advanced strategies and wire up Webhooks for automation.