space bunny

Overview

Space Bunny is an anonymous reasoning model presented through a free preview playground and a server-side API. It is designed for long-context work such as code review, research, visual understanding, and agent workflows.

Key Features

• A one-million-token context window for long documents, repositories, and extended conversations.
• Text, image, and compatible video URL inputs, with text responses for visual and textual questions.
• Five reasoning-effort levels: low, medium, high, xhigh, and max.
• JSON object responses for predictable downstream processing.
• Tool calling through approved functions, with the application retaining permission and side-effect control.
• OpenAI-compatible chat-completions requests, streaming, standard usage metadata, and server-side API key handling.

How It Works

1. Start with a prompt in the public playground, or create a Space Bunny API key for a server-side integration.
2. Send a chat-completions request with the model, messages, and the reasoning effort appropriate to the task.
3. Add image or video URL content when the request needs multimodal context, or request JSON when code needs structured output.
4. Read the assistant message and usage data, or validate any requested tool call before your application executes it.

Use Cases

• Review source code, architecture notes, and incident material across a large working context.
• Inspect screenshots, diagrams, or compatible video alongside an instruction.
• Build an agent that can request approved read-only or business functions while your application enforces permissions.

Pricing and Availability

The playground preview is free. The pricing page lists one-time credit packs with no subscription, no auto-renewal, and credits that do not expire; paid plans include Starter, Pro, and Enterprise options. The model is described as an anonymous preview, and the documentation says provider compatibility should be verified before production use.

Screenshots