⌁ Guides

oh-my-pi

Register Gondola as a provider in oh-my-pi's models.yml and run any Venice model from the omp TUI, paid per request in USDC on Base.

Create an API key

Sign in with your Base wallet at /keys and mint a Gondola API key. Keys start with gnd_. It authenticates every request and each call is charged in USDC against your Gondola balance.

Top up USDC on Base once at /wallet. The same balance powers the browser chat and every tool you connect.

Add Gondola to models.yml

oh-my-pi (omp for short) reads custom providers from ~/.omp/agent/models.yml. Add a Gondola block with explicit model entries:

~/.omp/agent/models.yml

# ~/.omp/agent/models.yml
providers:
  gondola:
    baseUrl: https://api.gondola-ai.com/v1
    apiKey: GONDOLA_API_KEY
    api: openai-completions
    authHeader: true
    models:
      - id: claude-sonnet-5
        name: Claude Sonnet 5
        contextWindow: 1000000
        maxTokens: 64000
      - id: qwen3-coder-480b-a35b-instruct-turbo
        name: Qwen3 Coder 480B
        contextWindow: 256000
        maxTokens: 65536

Export the key and launch:

shell

export GONDOLA_API_KEY=gnd_xxxx
omp

How the key is resolved

apiKey names an environment variable; if no variable with that name exists, omp uses the literal string as the token. authHeader: true is what injects it as the Authorization: Bearer header, so keep both lines.

Pick a model

Select Gondola models from the picker or reference them as gondola/<id>. Any id from the live catalog can be added to the models list the same way; copy context_length and max_completion_tokens into contextWindow and maxTokens:

shell

curl https://api.gondola-ai.com/v1/models

Why explicit entries instead of discovery

omp's discovery: type: openai-models-list also works against Gondola, but the catalog mixes text, image and video models, and discovery would list ids that cannot chat. Explicit entries keep the picker clean.

Prompt privacy

Gondola never stores prompts or completions: requests are metered, forwarded, and gone when the response finishes (the one exception, a video source clip over 18 MiB, is held privately for the length of its job; see /privacy), and Venice runs a no-logging model underneath. For zero retention on Venice-controlled infrastructure too, pick a model whose privacy field reads private on GET /v1/models. Details on /privacy.

Notes and troubleshooting

omp already ships Venice as a built-in provider. The Gondola entry serves the same catalog, routed per request within a small price band of the cheapest available supplier, below Venice list price, paid from one USDC balance, no Venice account or DIEM staking needed.

Cap the price on unattended runs

Agent sessions can run many steps unwatched. Add a provider-level header block with x-gondola-max-factor set to a value from 0 to 1 and Gondola refuses any request priced above that ceiling instead of paying more; omit it and a request with no cheaper capacity can fall through to Gondola's own backup key at Venice list price. For a hard dollar limit rather than a relative one, send x-gondola-max-charge-usd (for example 0.50) and any request whose estimated charge would exceed it is refused with a 402 before anything is billed.

If a request returns insufficient balance, top up at /wallet. Real usage and spend per key live at /keys.

Loading page

# ~/.omp/agent/models.yml providers: gondola: baseUrl: https://api.gondola-ai.com/v1 apiKey: GONDOLA_API_KEY api: openai-completions authHeader: true models: - id: claude-sonnet-5 name: Claude Sonnet 5 contextWindow: 1000000 maxTokens: 64000 - id: qwen3-coder-480b-a35b-instruct-turbo name: Qwen3 Coder 480B contextWindow: 256000 maxTokens: 65536