A standalone plugin and native provider bridge for OpenClaude that unlocks full access to OpenCode Go subscription models (https://opencode.ai/zen/go/v1).

Features:

Authentic OpenCode v2 Harness Reverse Engineering: Generates authentic 30-character descending session identifiers (ses_<timestamp_base62>), client affinity tracking, and full harness request headers (User-Agent, x-opencode-client, x-opencode-session, x-opencode-project, x-session-affinity).
Seamless Protocol Translation: OpenClaude talks standard OpenAI /v1/chat/completions. Upstream models like muse-spark-1.3-contributor require OpenAI /responses API. The bridge transparently translates requests and bi-directionally streams Server-Sent Events (SSE).
1M Native Context Window: Automatically configures OpenClaude for the full 1,048,576 token context window on muse-spark-1.3-contributor instead of falling back to 128K.
Configurable Reasoning Effort: Unlocks OpenClaude's native /effort selector (low, medium, high, xhigh) and translates effort parameters into upstream OpenCode Go reasoning structures.
Target Default Model: Defaults to muse-spark-1.3-contributor with verified support for space-bunny and all 36+ OpenCode Go models.
Model Name Normalizer: Strips any client prefixes (opencode-unlocked-*, opencode-go-*) to prevent upstream rejection.
Zero-Secret Distribution: The repository contains no hardcoded keys or .env files. Users provide their key via environment variable, interactive command, or local config.
Marketplace Ready: Fully compatible with OpenClaude's plugin and marketplace system (marketplace.json + plugin.json).

Quick Start

1. Installation

Option A: OpenClaude Marketplace

In OpenClaude, add the repository marketplace and install the plugin:

/plugin marketplace add /path/to/opencode-4-openclaude
/plugin install opencode-unlocked@opencode-unlocked

Option B: Direct Plugin Flag

Launch OpenClaude directly with the plugin:

openclaude --plugin-dir /path/to/opencode-4-openclaude

2. Configure Your API Key

Provide your OpenCode Go API key (oc_sk_...) using any of the following methods:

CLI Command:

./bin/opencode-unlocked set-key oc_sk_your_key_here

Environment Variable:

export OPENCODE_UNLOCKED_API_KEY="oc_sk_your_key_here"
# or
export GO_KEY="oc_sk_your_key_here"

3. Provider Setup for OpenClaude

Run the setup utility to configure OpenClaude's profile:

node bin/setup.mjs

This automatically backs up your existing profile and sets:

OPENAI_BASE_URL: http://127.0.0.1:8788/v1
OPENAI_MODEL: muse-spark-1.3-contributor

To restore your previous profile at any time:

node bin/setup.mjs restore

Daemon & CLI Management

Manage the background bridge daemon using ./bin/opencode-unlocked:

# Check status, active model, and token consumption
./bin/opencode-unlocked status

# Start daemon
./bin/opencode-unlocked start

# Stop daemon
./bin/opencode-unlocked stop

# Restart daemon
./bin/opencode-unlocked restart

# List all 36+ available models
./bin/opencode-unlocked models

# Switch active default model
./bin/opencode-unlocked switch space-bunny

# View or set reasoning effort (low, medium, high, xhigh)
./bin/opencode-unlocked effort high

# Run live connectivity verification
./bin/opencode-unlocked test

1M Context Window & Reasoning Effort

1,048,576 Token Context Window

OpenCode Go's muse-spark-1.3-contributor natively supports a 1M (1,048,576 tokens) context window. Normally, OpenClaude falls back to a 128,000 token limit on unknown OpenAI-compatible endpoints.

Opencode Unlocked automatically configures:

OpenClaude provider limits (maxContextLength: 1048576)
CLAUDE_CODE_OPENAI_CONTEXT_WINDOWS environment mapping
OpenClaude modelLimits in settings.json

Multi-Level Reasoning Effort

OpenCode Go /responses models support four levels of reasoning: low, medium, high, and xhigh.

You can adjust reasoning effort in two ways:

In OpenClaude Interactively: Use the native /effort slash command or the interactive picker to select low, medium, high, or xhigh.
Via Plugin CLI:
```
./bin/opencode-unlocked effort medium
```

Supported Models

All 36+ OpenCode Go models are available, including:

Model ID	Protocol	Specialization
`muse-spark-1.3-contributor`	`/responses` (Translated)	Target Default: Deep Reasoning & Coding
`space-bunny`	`/chat/completions` (Native)	Fast General & Code
`gpt-5.6-luna`	`/responses` (Translated)	High-Capability Reasoning
`deepseek-v4-flash`	`/chat/completions` (Native)	High-Throughput Flash
`glm-5.3`	`/chat/completions` (Native)	General Language
`kimi-k2.5`	`/chat/completions` (Native)	Long Context
`minimax-01`	`/chat/completions` (Native)	Conversational
`qwen3.5-max`	`/chat/completions` (Native)	General Capabilities

Architecture

OpenClaude CLI
      │ (OpenAI /v1/chat/completions)
      ▼
Opencode Unlocked Gateway (127.0.0.1:8788)
 ├── Harness Fingerprint (ses_ + descending base62 session ID)
 ├── Session Affinity Cache (prompt caching preservation)
 ├── Protocol Translation Engine
 │     ├── muse-spark-1.3-contributor -> POST /responses
 │     └── space-bunny / others       -> POST /chat/completions
 └── Dynamic Model Catalog & Prefix Normalizer
      │
      ▼
OpenCode Go Cloudflare Edge (https://opencode.ai/zen/go/v1)

Testing

Run the test suite:

# Unit & integration tests
npm test

# Live model verification (requires API key)
npm run test:live

License

MIT