Coding agents

Aider

Use Aider with an OpenAI-compatible provider.

Success — tested

2026-10-10. Windows text smoke test against a local fake gateway. Aider 0.86.2.

Aider works with code from your terminal. Use its OpenAI-compatible provider to reach POST /v1/chat/completions.

Before you start

Complete the Quickstart: add credits, create a key, and copy a model ID from Marketplace. Choose a model with tool support for agent tasks. The example ID is claude-sonnet-5-5; replace it with your selection. Install Aider and open your repository in a terminal.

Configure the tool

The /v1 suffix belongs in OPENAI_API_BASE. Aider's openai/ prefix selects its client adapter; it is not part of the gateway model ID.

macOS / Linux
export UNUSED_API_KEY='sk-unused-...'
export OPENAI_API_KEY="$UNUSED_API_KEY"
export OPENAI_API_BASE='https://unused.market/v1'
aider --model openai/claude-sonnet-5-5

For persistent model selection, put the following in .aider.conf.yml at your project root on any operating system. A home-wide file can also live at ~/.aider.conf.yml or %USERPROFILE%\.aider.conf.yml.

.aider.conf.yml
model: openai/claude-sonnet-5-5

Verify the connection

Ask: “Say hello in one sentence without using tools.” Then open Analytics → Requests. Find the new request by time and model. A response and a matching row confirm that traffic reached your account. This checks the connection, not the quality of repository edits. Start in /ask mode for a connection check before requesting edits.

Limits and troubleshooting

Aider may not recognize a canonical marketplace ID in its own model catalog. An unknown-model warning does not change the gateway ID; consult Aider's model settings before selecting an editing format. A text smoke test does not validate patch application.

SymptomFix
Authentication errorRe-enter your unused.market key.
No usable offerCheck Marketplace, Routing, and the key's allowed models.
Credit or budget errorCheck Credits and the key's monthly limit; reduce the output budget if needed.
Stream stops after output startsFind the failure in Requests. A started stream is not retried on another offer.

Sources and verification

Version checked: 0.86.2, 2026-10-10. Official sources: OpenAI-compatible APIs, configuration paths, model warnings.

On this page