Pricing
One-time license, runs on your machine. Bring your own model to keep monthly cost at zero, or wire up a managed provider when you want.
Early access: closed beta is free for invited users. Pricing shown is what you'll see at paid launch — sign in to lock in the early-access slot.
Local AI workstation for one person.
- Full desktop app (Windows / macOS / Linux)
- Pair web + mobile to your desktop (3 devices)
- Bring your own model — OpenAI, Anthropic, Ollama, Groq, etc.
- Local STT + TTS bundled
- All content + image generation tools
- Lifetime updates within v1.x
For power users + early-access channel.
- Everything in Standard
- 5 paired devices (vs 3)
- Early-access channel for v1.x features
- Priority email support during beta
- Public bug-tracker access
Common questions
What does "bring your own model" mean?
TALKY connects to any OpenAI-compatible API. That includes OpenAI itself, Anthropic via a proxy, Groq, DeepInfra, OpenRouter, and any local model server (Ollama, LM Studio, vLLM, llama.cpp). You pay your provider directly; TALKY takes no margin.
Is there a cloud / hosted version?
Not in v1. Your desktop is the source of truth; the web + mobile companion apps reach your desktop over an encrypted relay tunnel. A managed cloud tier is on the roadmap if demand warrants — feedback during beta drives the call.
What happens after v1?
Your license includes lifetime updates within the v1.x series. v2 (likely 12+ months out) will be a new license; v1 license holders get an upgrade discount.
Refund policy?
30-day no-questions refund once paid licenses open. During beta there's nothing to refund — it's free.