Build with Space Bunny Alpha

Quickstart

Send your first OpenRouter request

Start in the Playground, then move the same chat-completion shape behind your server with an OpenRouter API key.

1. 1

   Open the playground

   Sign in and test a real prompt with low reasoning effort before adding more context.

2. 2

   Create an OpenRouter key

   Store the key in a server-side environment variable. Never expose it in browser code.

3. 3

   Send chat messages

   Use model stealth/space-bunny-alpha with a standard system, user and assistant message history.

4. 4

   Add advanced controls

   Introduce reasoning effort, multimodal parts, tools or JSON output only when the workflow needs them.

Input modalities

Mix text with visual context

A user message can be a text string or an ordered content array. Put the text instruction first, then append image or video URL parts.

Space Bunny Alpha returns text. Video URL compatibility can vary by provider, so verify the exact route before production use.

Request controls

Control depth, modalities and tools

Keep the core request small: choose the model, provide messages, then add only the controls needed by the current task.

{
  "model": "stealth/space-bunny-alpha",
  "messages": [
    { "role": "user", "content": "Review this API design." }
  ],
  "reasoning": { "effort": "low" }
}

{
  "messages": [
    { "role": "user", "content": "Return a JSON launch plan." }
  ],
  "response_format": { "type": "json_object" }
}

{
  "model": "stealth/space-bunny-alpha",
  "messages": [{ "role": "user", "content": "Find the safest migration plan." }],
  "reasoning": { "effort": "high" }
}

{
  "role": "user",
  "content": [
    { "type": "text", "text": "Review this interface hierarchy." },
    { "type": "image_url", "image_url": { "url": "https://example.com/ui.png" } }
  ]
}

{
  "tools": [{
    "type": "function",
    "function": {
      "name": "get_record",
      "description": "Fetch one approved record",
      "parameters": { "type": "object", "properties": { "id": { "type": "string" } }, "required": ["id"] }
    }
  }],
  "tool_choice": "auto"
}

Output

Read the assistant message and usage

A successful non-streaming response contains choices. The first choice normally carries the assistant message, while usage reports prompt, completion and total tokens.

• message: Read choices[0].message.content for text or JSON, and message.tool_calls when the model requests a function.
• usage: Includes prompt_tokens, completion_tokens and total_tokens; OpenRouter may also include cost details.
• elapsedMs: Space Bunny's proxy adds end-to-end request time in milliseconds for Playground responses.

Reasoning tokens can count toward completion usage even when internal reasoning is not shown in the final message.

{
  "id": "chatcmpl-123",
  "model": "stealth/space-bunny-alpha",
  "choices": [{
    "index": 0,
    "message": { "role": "assistant", "content": "Here is the migration plan..." },
    "finish_reason": "stop"
  }],
  "usage": { "prompt_tokens": 161, "completion_tokens": 42, "total_tokens": 203 }
}

{
  "role": "assistant",
  "content": "The safest migration path is..."
}

{
  "role": "assistant",
  "content": "{\"risk\":\"low\",\"steps\":[\"backup\",\"migrate\"]}"
}

{
  "role": "assistant",
  "tool_calls": [{
    "id": "call_123",
    "type": "function",
    "function": { "name": "get_record", "arguments": "{\"id\":\"42\"}" }
  }]
}

API reference

Create a chat completion

Send OpenAI-compatible chat requests to OpenRouter with the Space Bunny Alpha model ID.

Authorization: Bearer <OPENROUTER_API_KEY>
Content-Type: application/json
HTTP-Referer: https://your-app.example
X-OpenRouter-Title: Your App

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stealth/space-bunny-alpha",
    "messages": [{ "role": "user", "content": "Review this design." }],
    "reasoning": { "effort": "low" },
    "max_completion_tokens": 2048
  }'

{
  "model": "stealth/space-bunny-alpha",
  "messages": [
    { "role": "system", "content": "Be careful and concise." },
    { "role": "user", "content": "Review this design." }
  ],
  "reasoning": { "effort": "low" },
  "temperature": 0.7,
  "max_completion_tokens": 2048
}

Keep the OpenRouter key in a server-side secret. This project's Playground calls its own server route so the browser never receives the provider key.

Agent usage

Use Space Bunny Alpha in an agent

Point an OpenAI-compatible client at OpenRouter, provide only approved tools, and keep validation and execution authority in your application.

export OPENAI_BASE_URL="https://openrouter.ai/api/v1"
export OPENROUTER_API_KEY="sk-or-v1-..."
export OPENAI_MODEL="stealth/space-bunny-alpha"

Review the supplied architecture notes, source files and incident logs. Identify the most likely failure boundary, cite the evidence for each conclusion, and propose the smallest safe fix. State what remains uncertain.

Use the available read-only tools to find the relevant customer record. Do not call destructive or payment tools. If a write would help, explain the proposed action and wait for explicit approval.

Analyze this migration plan for hidden rollback risks. Separate certain findings from assumptions, then recommend whether low, medium or high reasoning effort is appropriate for the final review.

Review the attached interface screenshot. Identify the three highest-impact hierarchy or accessibility problems and propose concrete fixes without changing the product's visual identity.

Return one valid JSON object with keys summary, risks, steps and verification. Do not use Markdown. Each step must include an owner and a measurable completion check.

Error handling

Errors and retries

OpenRouter uses standard HTTP status codes and returns a JSON error body when a request fails.

Retry 429 and transient 5xx errors with exponential backoff and jitter. Do not automatically retry invalid 400-series requests except rate limits.