Developer Tools
Cupid
Last updated 2026-10-07 · benchmark measured 2026-10-30 — deterministic & reproducible
AI coding CLI that routes each task step to the cheapest suitable model automatically.
Cupid at a glance
Cupid is a command-line tool for macOS, Linux, and Windows that breaks coding tasks into steps and automatically routes each step to the cheapest AI model (Claude, ChatGPT, or Gemini) that meets quality and risk requirements, for software developers and teams who want to control AI coding costs.
Good for
- Developers who want to reduce spending on AI coding assistants without choosing models manually
- Teams using Claude Code, Codex, or OpenCode who want automatic per-task model selection
- Cost-conscious developers who don't want to manage separate API keys for multiple AI providers
Pricing
Prepaid credits: $20/$50/$100 packs or custom $10-$1,000; billed at model cost + 5.5% fee, capped at an Opus-cost estimate. No subscription; online checkout not yet open.
Questions people ask
Does Cupid train on my code?
No, Cupid does not collect input to train models; router learning and training-capable models are both opt-in and off by default, and you can check or disable them with the cupid data status command.
Which tools does Cupid work with?
It works inside Claude Code, Codex, and OpenCode on macOS and Linux through one install command, plus the standalone Cupid CLI on macOS, Linux, and Windows.
Do I need my own API keys or a Claude subscription to use Cupid?
No, you sign in with a Cupid account and pay from one prepaid credit balance; provider API keys stay on Cupid's gateway and are not needed.
How is each request priced, and is there a limit on cost?
Each request is billed at the actual model cost plus a 5.5% fee, but the charge never exceeds Cupid's estimate of what that request would cost using Opus alone.
What happens if my credit balance runs out mid-request?
Cupid stops the request, only bills for work already completed (within the per-request cap), and lists the unfinished portion as remaining work; a request stopped due to insufficient balance is not charged.
Written by Legit.Show from cupidcode.ai on 2026-11-02. It does not change the score.
Are you the maker of Cupid? Claim this listing. It is free, takes a meta tag, a DNS record or GitHub admin rights, and never changes the score.
Launched something? Add your product, free.
To cite this score: legit.show/s/cupidcode-ai/2026-10-30. That address never changes; this page moves with every re-measure.
Legit.Show scored Cupid 96/100 on 2026-10-30, measured across all 7 frames from its public surface. legit.show/s/cupidcode-ai/2026-10-30
Is Cupid production-ready?
Legit.Show scores Cupid 96 out of 100 — the simple average of its 7 measured frames. Legit.Show ran its deterministic 7-Frame production-readiness benchmark on Cupid (public-surface assessment), measured from the public surface with no LLM in the scoring path. Its strongest frame is Security; its weakest is Standards. Every frame it averages is published with its evidence on the Legit.Show listing.
The 7 Frames
- Performance — 87/100
- Accessibility — 96/100
- Security — 100/100How this compares across the catalogue: The Web Security Baseline
- Privacy — 100/100How this compares across the catalogue: The Privacy Gap
- Reliability — 100/100
- Standards — 86/100
- Discoverability — 100/100
What we measured
- Security headers present: CSP, HSTS, X-Frame-Options, X-Content-Type-Options, Referrer-Policy.
- Served over HTTPS with a valid certificate.
- Real Lighthouse performance run — 350 ms to first byte.
- Returns a proper 404 for unknown routes.
- 3 of 3 sampled routes reachable.
- Has a reachable privacy policy.
- Sets cookies / loads scripts with no consent prompt.
- Discoverable: structured data, sitemap, OpenGraph image, canonical URL.
Who it's for
Software developers · Engineers · Teams using AI for coding · Cost-conscious developers
Pricing
Prepaid credits; signup grant to try for free, no card required; plan details coming
Sources and updates
- Description
- Taken from cupidcode.ai's own website on 2026-10-07.
- Operator
- 주식회사 큐피드(CUPID Inc.), as named in cupidcode.ai's terms, privacy page or footer.
- Benchmark
- Measured by Legit.Show from the live site, the way any visitor sees it — with no access to its code or accounts. Same method for every product, and no AI decides the score. Last checked 2026-10-30.
Put together from public information, without Cupid's involvement. If anything here is wrong, tell us and a person will check it.
Visit Cupid → · Alternatives to Cupid → · How this was measured →
