WorkBuddy × AI Code With
WorkBuddy Third-Party API / Relay Setup Guide
Use AI Code With inside the WorkBuddy desktop client. Add a custom model and custom provider to use Claude, ChatGPT, Gemini, DeepSeek, GLM, Kimi, and Grok.
Install WorkBuddy on macOS or Windows first, then finish setup in Settings. Enter the key only in WorkBuddy, never on this page.
- Provider
- Claude
- Base URL
- https://api.aicodewith.ai
- API Key
- sk-xxxxxxxx
- Model ID
- claude-sonnet-4-6
Quick setup
Save this key for later terminals
This tool stores the key in its own settings. After you sign in, copy the key and paste it into the API Key field in the steps below.
sk-***Signed-out visitors only see a placeholder. Sign in and we fill in your key. Sign in · Get a key in the console
Copying only puts the command on the clipboard. You still need to run it in a terminal, or paste the key into the tool settings. How to get an API key
In WorkBuddy Settings, add a custom model and custom provider, then enter the Base URL, API key, and Model ID for that family. About 3 minutes.
Open Settings, add a custom model and custom provider, then enter the Base URL, API key, and Model ID for the selected family.
Open WorkBuddy Settings
Open WorkBuddy, click the avatar or account area in the lower left, then choose Settings.
Open WorkBuddy, click the avatar or account area in the lower left, then choose Settings. Choose a family and Model ID
Pick Claude, ChatGPT, Gemini, DeepSeek, GLM, Kimi, or Grok, then copy a Model ID. Paste that ID into the model name field in WorkBuddy.
Choose a Model ID
Choose a provider, then pick a Model ID. Paste that ID into the model name field in WorkBuddy.
Select a model to copy its Model ID and see the matching Base URL.
Selected
- Provider
- Claude
- Model ID
- claude-sonnet-5-5
- Base URL
- https://api.aicodewith.ai
This page lists the current catalog by family. Latest Model IDs follow the models page and pricing page.
View the full model listOpen Models, then choose Add custom model. Add a custom provider and fill in three values
Open Models → Add custom model, then add a custom provider. The dialog title says only OpenAI-compatible APIs are supported; Claude uses the same dialog. Enter the Base URL, API key, and Model ID for the selected family.
Quick copy
Provider
Claude
Base URL
https://api.aicodewith.ai
Model ID
claude-sonnet-5-5
In the provider section, add a new custom provider. The dialog title says only OpenAI-compatible APIs are supported; Claude uses the same dialog. Enter the Base URL, API key, and Model ID. Claude uses https://api.aicodewith.ai with no /v1. Save, then send a message
Save, switch to the model you just added, and send a test message. A normal reply confirms the connection — then match the call by time and model in AI Code With usage records.
Save, switch to the model you just added, and send a test message. A normal reply confirms the setup.
How does WorkBuddy connect to AI Code With?
WorkBuddy connects to third-party services with a custom model and a custom provider. Compared with the click-by-click steps, remember three facts:
- Custom model
- Settings → Models → Add custom model. Third-party services are not in the built-in list, so you add them by hand.
- Custom provider
- Create a provider while adding the model. The dialog title says only OpenAI-compatible APIs are supported; Claude uses the same dialog — enter the Base URL from this page.
- Three core values
- The Base URL changes by model family: Claude uses https://api.aicodewith.ai; ChatGPT / DeepSeek / GLM / Kimi / Grok use https://api.aicodewith.ai/v1; Gemini uses https://api.aicodewith.ai/gemini_cli. Then add the API key and Model ID.
Flow: WorkBuddy client → custom provider + custom model → AI Code With API. Everything is configured inside WorkBuddy. No files to edit.
Frequently asked questions
New to WorkBuddy?
WorkBuddy is a desktop AI client for macOS (arm64 / x64) and Windows (x64 / arm64).
WorkBuddy official website
Download the official installer for your computer from the website, finish installation, then open WorkBuddy from the app list.
Open the WorkBuddy websiteFull install guidance follows the WorkBuddy official site; this site does not host WorkBuddy installers or images.
Common error troubleshooting
The most frequent problems fall into 4 categories.
401 / Invalid API Key
Symptom: WorkBuddy cannot start a response or reports an authentication error.
Cause: Usually the API key is invalid, copied incompletely, or expired.
- Open the AI Code With API key page and copy the complete key
- Return to WorkBuddy model settings and paste it again (no extra spaces)
- Save, then send a test message
Cannot find Add model
Symptom: Settings does not show a way to add a model.
Cause: You have not reached WorkBuddy's custom model management area, or you are in another Settings section.
- Open WorkBuddy and open the account menu in the lower left
- Choose Settings
- Open Models, then look for Add custom model
Model missing after saving / Model Not Found
Symptom: The model does not appear after saving, or the request reports Model Not Found.
Cause: The configuration was not saved, or the model name does not match the Model ID in the catalog.
- Open model settings and confirm the name matches the Model ID on this page exactly
- Save the configuration
- Look for the new item in the model list again
Base URL error / connection failed
Symptom: The request does not return a result after you enter the Base URL.
Cause: The Base URL is incomplete, or the endpoint does not match the model family.
- Claude uses https://api.aicodewith.ai (no /v1)
- ChatGPT / DeepSeek / GLM / Kimi / Grok use https://api.aicodewith.ai/v1
- Gemini uses https://api.aicodewith.ai/gemini_cli
Need the full guide?
Every config parameter, the complete model list, and more troubleshooting live in the AI Code With docs.
Open the AI Code With docs