spacebunny.me · Independent coding playground

Spacebunny

Paste a stack trace to fix a TypeScript error, or AI-summarize an RFC — with live credit metering so you do not burn an OpenRouter key. 100k context / request. Powered by DeepSeek V4 Flash via OpenRouter.

Not affiliated with DeepSeek, OpenRouter, or any model vendor. Buy credits when a demo proves useful.

Two workflows worth opening

Guides for the jobs people actually search — then jump into the playground demos.

Paste stack trace AI

Fix a TypeScript error with AI — stack first, minimal patch, credit estimate before send.

Read guide →

AI summarize RFC

Long document summary with AI credits under control — exec bullets, risks, and actions.

Read guide →

Built for one job: finish a real coding task

Not another raw proxy — metered playground with demos, templates, history, and export.

Finish a coding task, not a setup ritual

Drop a stack trace or diff, use a ready template, get a fix — no OpenRouter key wiring.

Long docs with a hard spend cap (100k tokens per request)

Paste large documents in one session. Soft warnings and a hard per-request cap keep bills predictable.

Credits you can see before you pay

Live token → credit estimates every turn. Prefer the $10 one-time pack until weekly use sticks.

Screenshot → answer

Upload UI shots or diagrams, then export the thread as Markdown when you are done.

What a paid-ready session looks like

Live credit estimates, templates, and export — so you know the cost before you top up.

deepseek/deepseek-v4-flash-0731 · temp 0.7 · 100k context / request~1,240 tokens · ~13 credits

Explain this React error boundary and suggest a minimal fix for streaming SSR.

An error boundary catches render errors in the child tree. For streaming SSR, wrap client islands that may throw, keep the boundary itself a client component, and avoid throwing during the shell render so the HTML stream can continue…

Next: summarize this 40k-token design doc — TOC first, then risks and open questions.

Structure the paste (index → priority sections → question) so the model spends attention where it matters and you stay under the per-request context cap…

Preview UI — sign in to save history, use prompt templates, and export Markdown.

Try free, then buy credits once

1 credit = 100 tokens (input + output). Preview models can vanish overnight — one-time credits are the safer default. Monthly plans only if you already use this weekly.

Free Tier

Free

100 credits

Finish a real coding or doc task — no card required

· 100 free credits on signup
· Chat with Spacebunny (DeepSeek V4 Flash)
· Prompt templates for code & docs
· Image upload + Markdown export

One-Time Credits

$10

1,500 credits

Pay once for coding / long-doc sessions — no subscription required

· 1,500 credits instantly
· No recurring charge
· Credits stay on your account
· Long context up to 100k tokens per request

Starter

$9 / month

1,200 credits / month

Monthly bucket if you use the playground every week

· 1,200 credits monthly
· 1-month rollover while subscribed
· Long context up to 100k tokens per request
· History, templates & export

Need more monthly volume? See Starter / Pro

Run one real task free

100 credits on signup — enough for a code review or a long-doc pass. If it helps, buy a $10 credit pack. No subscription required.