OpenCode

Use Dagu with OpenCode — a subscription service that gives you access to Kimi, DeepSeek, GLM, Qwen, MiMo, and other models through a single OpenAI-compatible API.

This applies to both:

• the Web UI steward
• workflow steps that use action: chat.completion or action: agent.run

What Is OpenCode

OpenCode (opencode.ai) provides a unified OpenAI-compatible endpoint for a rotating catalog of models. Access is subscription-based — you pay a flat monthly rate rather than per-token.

Base URL: https://opencode.ai/zen/go/v1

Environment variable: OPENCODE_API_KEY

Available Models

Model	ID
Kimi K2.6	kimi-k2.6
Kimi K2.5	kimi-k2.5
DeepSeek V4 Pro	deepseek-v4-pro
DeepSeek V4 Flash	deepseek-v4-flash
GLM-5.1	glm-5.1
GLM-5	glm-5
Qwen3.6 Plus	qwen3.6-plus
MiMo V2.5 Pro	mimo-v2.5-pro
MiMo V2.5	mimo-v2.5

The model catalog may expand over time. Check opencode.ai/docs/go for the current list.

Web UI Steward Setup

1. Go to /agent-settings
2. Enable Steward if not already on
3. Click Add Model
4. Fill in:
   • Provider: OpenCode
   • Model: e.g. kimi-k2.6
   • API Key: your OpenCode API key
   • Base URL: leave empty (defaults to https://opencode.ai/zen/go/v1)
5. Enable Supports Thinking if the model supports it (Kimi, DeepSeek V4 Pro, GLM, Qwen)
6. Click the star to set it as default

You can also pick from the built-in presets when adding a model — OpenCode presets are listed for all supported models.

Workflow Steps

Chat Step

steps:
  - action: chat.completion
    with:
      provider: opencode
      model: kimi-k2.6
      messages:
        - role: user
          content: "Summarize today's error logs."
    output: SUMMARY

The OPENCODE_API_KEY environment variable is read automatically. To use a different variable name:

steps:
  - action: chat.completion
    with:
      provider: opencode
      model: kimi-k2.6
      api_key_name: MY_OPENCODE_KEY
      messages:
        - role: user
          content: "Hello!"

With Thinking Enabled

Models that support reasoning (Kimi K2, DeepSeek V4 Pro, GLM-5, Qwen3):

steps:
  - action: chat.completion
    with:
      provider: opencode
      model: kimi-k2.6
      thinking:
        enabled: true
        effort: high
      messages:
        - role: user
          content: "Analyze the security implications of this code..."
    output: ANALYSIS

Agent Step

steps:
  - id: fix_config
    action: agent.run
    with:
      model: opencode-kimi-k2-6
      messages:
        - role: user
          content: "Review and fix the invalid entries in /etc/app/config.yaml"
    output: RESULT

The model field here refers to the saved model ID from /agent-settings.

Model Fallback

Use OpenCode as a fallback alongside other providers:

llm:
  model:
    - provider: anthropic
      name: claude-sonnet-4-6
    - provider: opencode
      name: kimi-k2.6
    - provider: opencode
      name: deepseek-v4-pro

steps:
  - action: chat.completion
    with:
      messages:
        - role: user
          content: "Explain this error."

Reasoning and Tool Calls

OpenCode reasoning models return thinking text in a reasoning field. Dagu preserves this between tool call turns, which is required by models like Kimi K2 when thinking is enabled. This is handled automatically — no extra configuration needed.

Related Pages

• Basic Chat
• Local AI
• Steward
• Models & Providers