Skip to main content

Zed Editor Integration Guide

Zed is a high-performance modern code editor with built-in AI assistant capabilities that supports multiple large language models. This guide explains how to integrate Zed Editor with OPE.AI to leverage cutting-edge models for programming assistance.

Quick Start​

1. Obtain API Key​

Visit the OPE.AI Console and create your dedicated API Key.

2. Add Provider​

You can configure OPE.AI using one of the following methods:

Add configuration through the Agent Panel visual interface without manually editing JSON files.

Steps​

  1. Press Cmd + Shift + A (Windows/Linux: Ctrl + Shift + A) to open the Agent Panel.
  2. Click + Add Provider in the LLM Providers area.
  3. Enter the configuration information according to the table below:
No.Field NameConfiguration Value
①Provider NameOPE.AI
②API URLhttps://api.ope.ai/v1
③API KeyYour OPE.AI API Key
④Model NameClaude-4.6-Sonnet
⑤Max Completion Tokens200000
⑥CapabilitiesCheck Tools and Images (model-dependent)
  1. Click + Add Model to continue adding other models (e.g., Claude-4.6-Opus, GPT-5.4-Pro).

Security Note: API Keys are securely stored in the system keychain (macOS Keychain / Linux Secret Service) and are not written to configuration files in plain text.


Getting Started​

Zed Agent Panel provides two working modes:

ModeDescription
Zed AgentBuilt-in AI assistant that uses your configured LLM Provider (e.g., OPE.AI).
External AgentsExternal integration plugins (e.g., Claude Code, Gemini CLI), running independently.

Start Zed Agent conversation:

StepOperation
①Click the + in the left panel → Select Zed Agent (shortcut: Cmd/Ctrl + N).
②If external plugins are installed, they will appear in the External Agents area.
③In the model selector at the bottom right, switch to the OPE.AI group and select a specific model.

Use CaseRecommended ModelFeatures
Core CodingClaude-4.6-Sonnet200K context, strong overall capability, fast response.
Top-tier ReasoningClaude-4.6-Opus200K context, strongest reasoning, ideal for architecture design.
Quick Q&AClaude-4.5-Haiku200K context, ultra-fast response, low cost.
Ultra-long ContextGPT-5.4-Pro1M context, OpenAI flagship model.
Cost OptimizationDeepSeek-V3.2128K context, extremely low cost, preferred for high-frequency calls.

For the complete model list, please check Model Pricing.


Model Configuration Reference​

Each model requires the following parameters:

FieldRequiredDescription
nameYesModel ID (e.g., Claude-4.6-Opus, GPT-5.4)
display_nameNoUI display name
max_tokensYesContext window size (e.g., 200000)
max_output_tokensNoMaximum output tokens (e.g., 64000)
capabilities.toolsNoWhether Function Calling is supported
capabilities.imagesNoWhether image input is supported

Common Model Configuration Quick Reference​

Model IDMax TokensMax OutputToolsImages
Claude-4.6-Opus20000064000YesYes
Claude-4.6-Sonnet20000064000YesYes
Claude-4.5-Haiku20000064000YesYes
GPT-5.4-Pro1050000128000YesNo
GPT-5.41050000128000YesNo
DeepSeek-V3.21280008000YesNo
Gemini-3-Pro100000064000YesYes

Troubleshooting​

OPE.AI Provider Not Showing in Model List​

  • Confirm that the Provider has been correctly added in the Agent Panel or settings.json.
  • Try restarting Zed Editor.
  • Check if the API URL is correct: https://api.ope.ai/v1

Authentication Error​

  • Confirm API Key format is correct (starts with sk-).
  • Open command palette Cmd/Ctrl + Shift + P → Search for language model: reset credentials → Re-enter Key.
  • Visit OPE.AI Console to confirm Key validity.

Tool Use Not Responding​

  • Confirm that "capabilities": { "tools": true } is set in the configuration.
  • Some models do not support tool calls; refer to the quick reference table above.

Model Returns Error or Timeout​

  • Check network connection, confirm access to https://api.ope.ai.
  • Try switching to another model (e.g., Claude-4.5-Haiku) for testing.
  • View Zed logs: Help > View Logs.