Free
$0
Download desktop for unlimited local inference on your hardware (Ollama, LM Studio, vLLM, llama.cpp). Chats stay in your folder. Hosted confidential models require Standard or Pro.
Pricing
Local is free forever on desktop. Hosted confidential AI is paid — Standard or Pro on desktop and the web. Standard: mid-model continuity + Standard search + Agent Teams Taste. Pro: Research and fuller Agent Teams. First paid period: 30-day money-back.
Desktop apps are available now (macOS Apple Silicon, Windows, Linux). Hosted Standard and Pro are sold as SaaS subscriptions. Global card checkout is invite-only while we finish payment onboarding — contact support@teechat.ai or use in-app billing when offered.
Most started
Standard
$10 / month
or $100 / year (2 months free)
Hosted mid open models (~31–35B) with Standard search and Agent Teams Taste (5/day) — desktop and web. Conversations stay in your folder on desktop — not our cloud. Research stays on Pro.
Pro
$20 / month
or $200 / year (2 months free)
Desktop and web. Higher quotas, metered Research with sources, Agent Teams Standard (10/day · 120/month), Pro scheduled-task limits, and priority for larger open models when we ship them.
Compare
| Free | Standard | Pro | |
|---|---|---|---|
| Clients | Desktop (local) | Desktop + web | Desktop + web |
| Local inference (your engine) | Yes | Yes | Yes |
| Hosted confidential models | No | Mid open models | Higher quotas + larger models when ready |
| Standard search | — | Yes | Yes |
| Research mode | — | — | Yes |
| Agent Teams | — | Taste · 5 / day | 10 / day · 120 / month |
| Hosted context | Your local stack | 64K | 256K engine window (not an SLA) |
| OpenAPI keys (drop-in tools) | — | Yes | Yes |
Hosted is pay-first: 30-day money-back on your first paid Standard or Pro subscription. No usage-cap refund denials. Local stays free forever.
Team / organization seats (3+) are available by application — ops reviews before seats activate.