Reuse, don't regenerate
Your AI reads what it already worked out — instead of spending tokens to figure the same thing out all over again.
Your AI saves what it already worked out and reuses it — so it burns far fewer tokens on every task. Host it for your whole team in a few clicks, or run it free on your own machine.
Few-click hosted setup for teams — or free on your own computer, no account needed.
Your AI reuses what it already worked out, instead of spending tokens to generate it again from scratch.
Your AI reads what it already worked out — instead of spending tokens to figure the same thing out all over again.
Big jobs usually burn tokens repeating themselves. Reusing earlier context keeps the token count low from start to finish.
Everyone's reused context adds up. Your dashboard shows the tokens — and the spend — you're saving company-wide.
Point Claude, Codex, or Cursor at Scholialang. Nothing about how you work changes — your AI just reuses context and spends far fewer tokens.
Hosted plans differ by how much you run, whether you're a team, and whether the cloud writes your traces for you. Prefer to self-host? Local-first stays free, forever.
Your traces, off your laptop.
$5/person/mo
Solo, just moving to the cloud.
For heavy solo use.
$12/person/mo
Individuals running lots of traces.
Teams + AI writing, offloaded.
$24/person/mo
Teams that need collaboration, admin & compliance.
Normally your machine runs the Gemma model to write and summarize every reasoning trace. On the upper plans a hosted Gemma4 model does it in our cloud — no local compute, faster, and consistent. Included on Business, pay-as-you-go on Pro.
| Compare plans | Freelocal | Starter$5/mo | Pro$12/mo | Business$24/mo |
|---|---|---|---|---|
| Local-first (SQLite, exports) | ✓ | ✓ | ✓ | ✓ |
| Hosted cloud trace storage | — | ✓ | ✓ | ✓ |
| Usage limits | — | Basic | Higher | Highest |
| Search across traces | — | ✓ | ✓ | ✓ |
| Official app relay | — | — | ✓ | ✓ |
| Retention window | Local | 7 days | 30 days | Configurable |
| Gemma4 Cloud — offloaded trace writing | — | — | Pay-as-you-go | Included |
| Team workspaces & shared projects | — | — | — | ✓ |
| Members, roles & seat admin | — | — | — | ✓ |
| SSO (SAML / OIDC) & SCIM | — | — | — | ✓ |
| Audit log & retention controls | — | — | — | ✓ |
| Dedicated storage / BYO-Cloudflare | — | — | — | ✓ |
| Support | Community | Community | Priority |
Every reasoning trace your agent writes to the cloud is one MCP trace write. The monthly allowance is set by your workspace tier and pooled across every seat — beyond it, writes are metered at $0.05 per 1,000.
The free plan keeps 1user on your own computer — it isn't a trial that runs out.
Get your team on hosted reasoning traces — or run it free on your own machine.