Chatbox × AI Code With
Use AI Code With in Chatbox: Third-Party API / Relay Setup
This page shows how to connect a third-party API in Chatbox, an open-source desktop client, so you can manage models and chats locally with a custom provider.
Download and install official Chatbox first, then add a custom provider in Settings; the API key is only entered in Chatbox, never on this page.
- API mode
- OpenAI API compatible
- API Host
- https://api.aicodewith.ai
- API Path
- /chat/completions
- API Key
- sk-xxxxxxxx
- Model ID
- deepseek-v4-flash
Quick setup
Install Chatbox from the official site, add a custom provider, and enter the API host without /v1 plus your API key. Use DeepSeek, Gemini, GLM, Kimi, and more in about 3 minutes.
Download and install official Chatbox first. Open Settings → Model Providers → + Add → Add custom provider. Set API mode to OpenAI API compatible, host https://api.aicodewith.ai (no /v1), and keep path /chat/completions.
Download and install official Chatbox
Open the Chatbox website and download the official installer for Windows, macOS, or Linux. This site does not host installers.
Open the Chatbox websiteAdd a custom provider and fill host plus key
Open Settings → Model Providers, click + Add, then Add custom provider. Name it AI Code With. Set API mode to OpenAI API compatible. Enter https://api.aicodewith.ai without /v1. Keep the default path /chat/completions — Chatbox inserts /v1 itself. Paste your API key, then Check if you want; a passing Check is not a successful connection.
Quick copy
API host (no /v1)
https://api.aicodewith.ai
API path (keep default)
/chat/completions
Model ID
deepseek-v4-flash
After copying, return to Chatbox: paste the API host into the custom provider and keep path /chat/completions. Paste the API key only inside the client. Do not paste Cherry Studio / Zed URLs that include /v1.
Open Chatbox Settings, go to Model Providers, and click + Add at the bottom of the list. Do not switch to a built-in OpenAI, Gemini, or Claude provider. At the bottom of the picker, choose Add custom provider. Do not import from clipboard, and do not pick built-in OpenAI or OpenRouter. Name it AI Code With. Set API mode to OpenAI API compatible. Copy the host from this page: https://api.aicodewith.ai with no /v1. Keep the default path /chat/completions — Chatbox inserts /v1 itself, so the preview should be https://api.aicodewith.ai/v1/chat/completions. Paste the API key, then Fetch models; if the list is empty, New and type the Model ID. A passing Check is not a successful connection. Fetch or create models and verify
In Models, click Fetch; if the list is empty, New and type the full Model ID. On the chat page, pick that custom provider and model, then send a test message. A normal reply means the setup works — then match the call by time and model in AI Code With usage records. Saving settings is not enough.
Copy a Model ID
After adding the provider, click Fetch, or New and enter one of these Model IDs:
Pick a family, then copy a Model ID. Fetch the list, or New and paste if it is empty.
Selected
- Provider
- DeepSeek
- Model ID
- deepseek-v4-flash
The full model list follows the AI Code With docs and pricing page; this site does not maintain a separate model list. Do not prefix IDs with openai/. Gemini models need Google Gemini API compatible mode on that model.
View the full model list
How does Chatbox connect to AI Code With?
Chatbox calls compatible APIs through a custom OpenAI-compatible provider. Connecting to AI Code With only needs Settings:
- Custom provider
- Open Settings → Model Providers, click + Add, then Add custom provider. Do not switch to built-in OpenAI or OpenRouter.
- API host
- Set API mode to OpenAI API compatible. Enter https://api.aicodewith.ai without /v1. Keep the default path /chat/completions; Chatbox inserts /v1 itself.
- API Key
- Paste your AI Code With API key. It stays on this device — do not put it in a public repo or a screenshot.
- Model ID
- Click Fetch to pull the list; if it is empty, New and type the full Model ID (for example, deepseek-v4-flash). For Gemini models, change that model's API mode to Google Gemini API compatible.
Flow: Chatbox → custom OpenAI-compatible provider → AI Code With API. Install the official client first, then add the provider in Settings. A passing Check is not a successful connection.
Frequently asked questions
Don't have Chatbox installed?
Download the official client for your OS from the Chatbox website. This site does not host installers.
Chatbox website
Open chatboxai.app and download the official installer for Windows, macOS, or Linux. Use the official client for this integration.
Open the Chatbox websiteFull install guidance follows the official Chatbox site. Do not paste other API hosts from third-party tutorials.
Common error troubleshooting
The most frequent problems fall into 4 categories.
401 / Invalid API Key
Symptom: Sending a message or clicking Check returns 401, Unauthorized, or Invalid API Key.
Cause: Usually the API key is invalid, mistyped, or copied with extra characters.
- Open the custom provider and re-paste the AI Code With API key (no extra spaces)
- Confirm the key is a valid AI Code With API key and the account has balance
- Do not put the key on this page or in a public repository
Model Not Found / Wrong Model ID
Symptom: The request reports Model Not Found, unknown model, or 404, or the new model is missing from chat.
Cause: The Model ID was mistyped, Gemini is still on OpenAI-compatible mode, or the model is currently unavailable.
- Copy the exact Model ID from this page (e.g. deepseek-v4-flash) instead of typing it
- For Gemini models: open that model's settings and switch API mode to Google Gemini API compatible
- Confirm the current key can access that model
API host error (usually an extra /v1)
Symptom: Sending a message fails, times out, or reports a wrong endpoint path / 404.
Cause: The API host includes /v1, the default API path was changed, or another relay domain was used.
- Confirm the host is https://api.aicodewith.ai (no /v1)
- Keep the path /chat/completions; the preview should be https://api.aicodewith.ai/v1/chat/completions
- Confirm you are using the custom provider you just added, not built-in OpenAI
Fetched model list is empty
Symptom: Fetch returns an empty list, or chat cannot pick an AI Code With model.
Cause: The host is unreachable, the key is invalid, or the upstream does not return a full list.
- Confirm host and key, then Fetch again
- If the list is still empty, click New and type the full Model ID
- Save, pick that model on the chat page, then send
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