Loading…
LUNAX · CATALOG
LUNAX · CATALOG
LUNAX · CATALOG
AI subscription
| Feature | Free | Plus | Pro | Max |
|---|---|---|---|---|
| Subscription tokens / month | 10,000,000 | 25,000,000 | 100,000,000 | 250,000,000 |
| Tokens / 5-hour window | Monthly limit only | 1,250,000 | 5,000,000 | 12,500,000 |
| Tokens / 7-day window | Monthly limit only | 6,250,000 | 25,000,000 | 62,500,000 |
| Total token ceiling with PAYG | 10,000,000 | 75,000,000 | 300,000,000 | 750,000,000 |
| Max turns per session | 12 | 24 | 40 | 60 |
| Max tool calls per turn | 50 | 100 | 180 | 250 |
| Frontier models access | Standard fast models | All frontier models (Claude, Qwen) | All models + priority execution queue | All models + dedicated compute pool |
| Free dedicated databases | PostgreSQL 16 HA + MySQL + Redis | PostgreSQL 16 HA + MySQL + Redis | PostgreSQL 16 HA + MySQL + Redis | PostgreSQL 16 HA + MySQL + Redis |
| Concurrent AI editor sandboxes | 0 | 1 | 3 | 10 |
| Included active starter app slots | 1 | 3 | 5 | 5 |
| Bounded included compute credit | $0.00 | $0.10 | $0.38 | $0.96 |
| Included runtime envelope | 512Mi · shared CPU · idle sleep · 1 replica | 512Mi · shared CPU · idle sleep · 1 replica | 512Mi · shared CPU · idle sleep · 1 replica | 512Mi · shared CPU · idle sleep · 1 replica |
| Host your code without AI subscription | Yes, within hosting limits | Yes, within hosting limits | Yes, within hosting limits | Yes, within hosting limits |
| Git tools in AI editor | Not included | Not included | Enabled | Enabled |
| AI browser automation | Not included | Not included | Enabled | Enabled |
| AI subagent tools | Not included | Enabled | Enabled | Enabled |
| AI database tools | Enabled | Enabled | Enabled | Enabled |
| Owned custom domain support | Yes, within hosting limits | Yes, within hosting limits | Yes, within hosting limits | Yes, within hosting limits |
Paid windows share the monthly pool: 25% per 7 days and 20% of that weekly budget per 5 hours. Each window opens on the first admitted prompt and resets after its full duration. They are not extra token grants.
Slots count logical active applications, not builds, redeploys or replicas. The approved counts are 1 / 3 / 5 / 5. Compute credit is separate and must be configured before these starter offers become available. MAX adds AI capacity, not unlimited servers.
Application hosting
| Feature | Pak · 512Mi | Nab · 1Gi | Khas · 2Gi | Bekr · 4Gi | Tak · 8Gi | Ouj · 16Gi |
|---|---|---|---|---|---|---|
| Billing term | 374,000 Toman / month 374,000 Toman billed every 1 months | 690,000 Toman / month 690,000 Toman billed every 1 months | 1,290,000 Toman / month 1,290,000 Toman billed every 1 months | 2,390,000 Toman / month 2,390,000 Toman billed every 1 months | 4,490,000 Toman / month 4,490,000 Toman billed every 1 months | 8,390,000 Toman / month 8,390,000 Toman billed every 1 months |
| AI database tools | 1 managed Postgres database + Redis included | 1 managed Postgres database + Redis included | 1 managed Postgres database + Redis included | 1 managed Postgres database + Redis included | 1 managed Postgres database + Redis included | 1 managed Postgres database + Redis included |
Sleeping HTTP apps wake on a request and may have a cold start. Bots and workers need always-on hosting; extra replicas and larger sizes require funded opt-in and an explicit spend cap. Shared CPU is not a dedicated core.
Log in to view your live token usage and plan limits.
Log in220,000 Toman per million extra tokens. Requires PAYG consent and a funded LIVE wallet.
Invitation: 14 days of PLUS with 100,000,000 to 1,000,000,000 gift tokens. Gift and purchased tokens are separate from subscription windows.
Trial expiry ends PLUS entitlement. Source, revisions and data are retained. New hosting work uses the current tier, configured budget and spending consent. No lifetime hosting is promised.
Cloud
No. The same logical production application retains one slot. Failed builds and release history do not count. Previews and replicas have separate resource/time limits.
New billable work is refused. PAYG is never enabled automatically. Source and data are retained; continued runtime suspension and notification require the reviewed release gate before paid hosting is offered.
No. Use the web Git path or the supported source-checkout CLI. Login, link, init, env set and deploy are real commands. The CLI package is private; no public installer is advertised.
No. Static artifacts serve files; a container executes code and consumes CPU/RAM while awake. Bots and background workers cannot rely on HTTP wake-up.
No. PLUS expires after the invitation period. Gift tokens are not hosting credit. Your code and data remain available under the current hosting policy.