Setup opencode

Setup Open Code Agent

Step 0: Language selection

Before proceeding with any other steps in this skill, ask the user which language to continue in using AskUserQuestion. Keep this initial prompt in English because the preferred language is not yet known.

• Question: Which language should I continue in?
• Options: English, 日本語 (Japanese)

Use the selected language for all subsequent user-facing messages and for every further AskUserQuestion prompt in this skill. Do not translate code, file paths, shell commands, or file contents.

Configure nagi to use Open Code instead of Claude Code. Open Code supports multiple AI providers (OpenRouter, Gemini, OpenAI).

UX Note: Use AskUserQuestion for all user-facing questions.

Step 1: Pre-flight

Check Docker

docker info > /dev/null 2>&1 && echo "RUNNING" || echo "NOT_RUNNING"

If not running, start Docker first.

Check if image exists

docker images nagi-agent-opencode --format "{{.Repository}}:{{.Tag}}" | head -1

If no image, it will be built in Step 5.

Step 2: Choose Provider

AskUserQuestion: Which AI provider do you want to use?

Provider	Best for	API Key Source
OpenRouter	Access to many models (Claude, GPT, Gemini, etc.) via one key	https://openrouter.ai/keys
Google Gemini	Google's Gemini models directly	https://aistudio.google.com/apikey
OpenAI	GPT-4o, o1, etc.	https://platform.openai.com/api-keys

Step 3: Get API Key

Based on the selected provider, guide the user:

OpenRouter

AskUserQuestion: Paste your OpenRouter API key

1. Go to https://openrouter.ai/keys
2. Create a new key
3. Copy the sk-or-... key

Google Gemini

AskUserQuestion: Paste your Google AI API key

1. Go to https://aistudio.google.com/apikey
2. Create API key
3. Copy the key

OpenAI

AskUserQuestion: Paste your OpenAI API key

1. Go to https://platform.openai.com/api-keys
2. Create new secret key
3. Copy the sk-... key

Step 4: Choose Model

AskUserQuestion: Which model do you want to use? (suggest defaults based on provider)

OpenRouter defaults:

• openrouter/anthropic/claude-sonnet-4 (Recommended)
• openrouter/openai/gpt-4o
• openrouter/google/gemini-2.5-pro

Google Gemini defaults:

• google/gemini-2.5-pro (Recommended)
• google/gemini-2.5-flash

OpenAI defaults:

• openai/gpt-4o (Recommended)
• openai/o1

Step 5: Configure .env

Add/update the following in .env:

# Open Code settings
CONTAINER_IMAGE=nagi-agent-opencode:latest
OPENCODE_MODEL={provider}/{model}

And the provider-specific API key:

OpenRouter

OPENROUTER_API_KEY=sk-or-...

Google Gemini

GOOGLE_API_KEY=...

OpenAI

OPENAI_API_KEY=sk-...

Step 6: Build Docker Image

./container/open-code/build.sh

This builds the nagi-agent-opencode:latest image with Open Code CLI.

Wait for build to complete, then verify:

docker images nagi-agent-opencode --format "{{.Repository}}:{{.Tag}} {{.Size}} {{.CreatedSince}}"

Step 7: Restart nagi

launchctl kickstart -k gui/$(id -u)/com.nagi.{ASSISTANT_NAME}
sleep 3
launchctl list | grep com.nagi.{ASSISTANT_NAME}
tail -5 __data/{ASSISTANT_NAME}/logs/nagi-{ASSISTANT_NAME}.log

Step 8: Test

Tell user:

Send a message in Slack/Discord. The agent should now respond using Open Code with your selected model.

Check logs if no response:

tail -20 __data/{ASSISTANT_NAME}/logs/nagi-{ASSISTANT_NAME}.log

Switching back to Claude Code

To switch back, update .env:

CONTAINER_IMAGE=nagi-agent:latest

Remove or comment out the Open Code settings:

# OPENCODE_MODEL=...
# OPENROUTER_API_KEY=...

Then restart nagi.

Troubleshooting

Container fails to start

• Check Docker is running: docker info
• Rebuild image: ./container/open-code/build.sh
• Check logs: tail -20 __data/{ASSISTANT_NAME}/logs/nagi-{ASSISTANT_NAME}.error.log

No response from agent

• Verify API key is correct
• Check CONTAINER_IMAGE=nagi-agent-opencode:latest is set in .env
• Check OPENCODE_MODEL is set correctly
• Try a different model if rate limited

Wrong model responding

• Verify OPENCODE_MODEL in .env matches your intended provider/model
• Restart nagi after changing .env