Models
Fast models by default.Your choice when you need it.
HiberChat runs on Google Gemini Flash-Lite for fast, affordable answers, and supports OpenAI, Groq, and any OpenAI-compatible provider.
10-day Pro trial · No credit card · Free plan forever
Model
Google Gemini
gemini-flash-lite-latest
OpenAI
any GPT model
Groq
openai/gpt-oss-20b
OpenAI-compatible
any endpoint
Latency
0.52s
Tokens in
1,204
Tokens out
86
What's included
Everything in Models & Providers.
Gemini Flash-Lite
The default model replies in roughly half a second without slow reasoning overhead.
OpenAI
Use GPT models with your own API key on self-hosted installs.
Groq
Low-latency open models such as gpt-oss-20b through Groq.
OpenAI-compatible
Point HiberChat at any compatible endpoint, including self-hosted models.
Keys stay server-side
Provider keys never reach the browser or the widget.
Usage tracked
Token counts and latency are recorded for every AI reply.
The right model for the job.
Speed matters in live chat. We tune for fast first replies.
Default
Fast and grounded.
Gemini Flash-Lite with retrieval from your content.
Flexible
Bring your provider.
Swap models with configuration, not code.
Safe
Fallback built in.
If a provider fails, an extractive answer from your sources is used.
Questions about Models & Providers
Do I need my own AI key?
Is there a free trial?
Scale customer chat. Not headcount.
An AI agent trained on your website, a shared inbox for your team, and every channel in one place.
Free forever for one agent. AI from $25 per agent.
10-day Pro trial, no credit card. Cancel or downgrade any time.