Integrations
Zed is an open-source code editor with a built-in AI agent. Add DEVUP AI as an OpenAI-compatible provider, and the Zed agent, the inline assistant and chat run on DEVUP AI models.
On Windows:
winget install -e --id ZedIndustries.ZedOn macOS and Linux, download Zed from https://zed.dev/download.
| Step | Where | What to do |
|---|---|---|
| 1 | Zed | Open Settings with Ctrl+,. |
| 2 | Settings, section AI | Open LLM Providers and click Add Provider. |
| 3 | Provider form | Provider Name: DEVUP AI. API URL: https://api.devupai.com/v1. API Key: your DEVUP AI key. |
| 4 | Models | Model Name: deepseek-ai/DeepSeek-V4-Pro. Max Tokens: the model's context_window (next section). Keep the values Zed fills in for Max Completion Tokens and Max Output Tokens. |
| 5 | Provider form | Save. |
https://api.devupai.com/v1
Zed stores the key in the system keychain, not in settings.json.
Max Tokens is the model's context window. Read it from context_window in GET /v1/models:
$models = Invoke-RestMethod -Uri "https://api.devupai.com/v1/models" -Headers @{ Authorization = "Bearer $env:DEVUP_API_KEY" }
$models.data | Where-Object { $_.id -eq "deepseek-ai/DeepSeek-V4-Pro" } | Select-Object id, context_windowEnter the number it prints for the model you added.
| Checkbox | Setting |
|---|---|
| Supports tools | On |
| Supports images | Off |
| Supports parallel_tool_calls | Off |
| Supports prompt_cache_key | Off |
| Supports /chat/completions | On |
| Uses max_tokens for output limit | Off |
| Supports thinking | Off |
Open the Agent Panel and pick the model under DEVUP AI.
Reply with exactly: Salam DEVUP
Open a project folder, keep the Write profile selected next to the model picker, and ask:
Create hello.go that prints Salam DEVUP, then read the file back and tell me its first line.
The agent creates the file, reads it back and answers. Review its changes with Keep All or Reject All.
Select code, press Ctrl+Enter, and type your instruction:
Add a comment explaining this function
Add one model entry per catalog ID you want in Zed, each with its own context_window. The agent needs a model with tool calling. See Models and Tool Calling.
Yes. Add an OpenAI-compatible provider in Zed's LLM Providers settings with the API URL https://api.devupai.com/v1 and your DEVUP AI key, then pick the model in the Agent Panel.
Yes. Keep Supports tools on and the Write profile selected, and the agent creates, reads and edits files in your project.
Every request from Zed is a regular DEVUP AI API call, metered in Algerian Dinar on your account.
In the system keychain. It is not written to settings.json.
The model's context window: the context_window value that GET /v1/models returns for that model.
Yes. Select code, press Ctrl+Enter and write your instruction; the inline assistant uses the DEVUP AI model you picked.