Opencode

Overview

Opencode exposes two OpenAI-compatible AI gateways that share a single DeepIntShield implementation, differing only in default base URL and provider key:

Provider Key	Gateway	Default Base URL	Billing
opencode-zen	Opencode Zen	https://opencode.ai/zen	Pay-as-you-go
opencode-go	Opencode Go	https://opencode.ai/zen/go	Subscription

Key features:

• Full OpenAI compatibility - Identical request/response format
• Streaming support - Server-Sent Events with delta-based updates
• Tool calling - Complete function definition and execution support
• Opencode-aware errors - Parses Opencode’s {"type":"error","error":{...}} envelope

Supported Operations

Operation	Non-Streaming	Streaming	Endpoint
Chat Completions	✅	✅	/v1/chat/completions
Responses API	✅	✅	/v1/chat/completions
List Models	✅	-	/v1/models
Text Completions	❌	❌	-
Embeddings	❌	❌	-
Image Generation	❌	❌	-
Speech (TTS)	❌	❌	-
Transcriptions (STT)	❌	❌	-
Files	❌	❌	-
Batch	❌	❌	-

Note

Unsupported Operations (❌): Text Completions, Embeddings, Image Generation, Speech, Transcriptions, Files, and Batch are not supported by the Opencode gateways. These return an unsupported-operation error.

---

Configuration

Add Opencode as a provider with a single API key. Choose the provider key for the gateway you subscribe to:

{
  "provider": "opencode-zen",
  "keys": [{ "value": "env.OPENCODE_API_KEY", "models": [], "weight": 1.0 }]
}

For the subscription gateway use "provider": "opencode-go". Both default base URLs can be overridden per workspace.

---

1. Chat Completions

Request Parameters

Opencode supports the standard OpenAI chat completion parameters. For full parameter reference and behavior, see OpenAI Chat Completions.

---

2. Responses API

The Responses API is converted internally to Chat Completions and the chat response is mapped back to the Responses shape.

---

3. List Models

Opencode’s model listing endpoint returns available models at /v1/models.

---

Error Handling

Opencode returns errors in a custom envelope rather than the OpenAI shape:

{ "type": "error", "error": { "type": "...", "message": "..." } }

DeepIntShield parses this envelope so the surfaced error message and type reflect Opencode’s response while preserving the HTTP status code.