Bots that actually know your product
Feed JestBot your docs, drop in a snippet, and it answers your customers’ questions — grounded in what you actually wrote, then handed off to a real agent when it matters.
No card required · Free plan included · Live in under two minutes
Acme Support
Powered by JestBot
< 2 min
Time to first answer
Your own docs
Grounded in
A real agent
Escalates to
$0
Starts at
Works with the model provider you already use
How it works
From your docs to a live chat widget in minutes
No vector database to stand up, no prompt to hand-tune — feed it content, embed a script tag, and JestBot takes it from there.
Feed it anything
Paste raw text, point it at a URL, or upload a PDF, DOCX, TXT, CSV, or Markdown file. Everything gets chunked automatically so answers stay grounded in real context.
- Text & URL ingestion
- PDF / DOCX / CSV / MD upload
- Automatic chunking
Grounded, not guessed
Replies are built from the chunks your bot actually retrieved for that question — visible with a similarity score in Test Chat, not a black box.
- Cosine-similarity retrieval
- Streaming replies over SSE
- Retrieved chunks shown for debugging
Bring your own model
Start free on the default Ollama-backed model, then switch any bot to your own OpenAI, Anthropic, or Gemini key the moment you need more.
- Free tier by default
- Per-bot provider selection
- Keys encrypted at rest
One script tag to go live
NewDrop a single embed snippet before </body> and the widget just works — themeable, no build step, no framework dependency.
- Zero-dependency widget
- Public key safe client-side
- Fully themeable
Beyond the bot
A real support inbox waiting behind the AI
Most questions never need a person. The few that do get handed off instantly, inside business hours, to a team that's actually available.
Human handover, on your terms
Set a message threshold — the AI answers first, then a "Talk to a human" option appears and hands the live conversation over with zero context lost.
- Configurable handover threshold
- Assign agents or whole teams
- No message loss on handoff
Business hours that visitors can see
NewInside hours, visitors see agents are available. Outside hours, the AI says so plainly, keeps chatting, and captures a follow-up email instead.
- Per-bot schedule & timezone
- Visible availability
- Off-hours email fallback
Agents, teams & performance
Give agents their own login and panel, group them into teams, and track status plus resolved, missed, and reassigned counts per person.
- Dedicated agent panel
- Team-based assignment
- Live performance tracking
CSAT on every resolution
NewWhen an agent resolves a chat, the visitor gets a quick 1–5 star prompt with an optional comment — rolled up per agent and per bot.
- 1–5 star rating
- Per-agent & per-bot rollups
- Prompted automatically
One script tag. That’s the whole integration.
Drop the embed snippet before </body> on any page — no build step, no framework dependency. Prefer a custom UI? Call the same chat endpoint directly and stream responses over SSE.
- Public key is safe to expose client-side — it’s read-only.
- Server-Sent Events stream tokens as they’re generated.
- Session IDs keep the last 10 messages of context automatically.
<script src="https://your-api-domain/widget.js?key=pk_..." async></script>A generic chatbot vs. one that’s read your docs
Live in three steps
01
Create a bot
Name it, describe what it should help with. Takes under a minute.
02
Add knowledge
Text, URLs, or files — JestBot chunks and embeds it all automatically.
03
Embed & go live
One script tag, and it's answering questions on your site immediately.
Built for teams that ship
Assign bots to agents or whole teams, keep every conversation logged for review, and scope who can touch what — without slowing anyone down.
AI first, human when it matters
Most questions never need a person. The ones that do get handed off instantly, inside business hours, to whoever’s actually available.
Built to be trusted with your content
Keys encrypted at rest
Bring your own OpenAI, Anthropic, or Gemini key and it's encrypted before it ever touches disk — never logged in plaintext.
Domain allowlisting
Restrict which sites are allowed to load your widget. Leave it empty during testing, lock it down before launch.
Delete anytime
Remove a single document or an entire bot — including its conversation history — whenever you want. Nothing lingers.
Builders shipping with JestBot
“We replaced a wall of FAQ text with a bot that actually points people to the right paragraph. Setup was one afternoon.”
Priya S.
Founder, early-stage SaaS
“Being able to see the retrieved chunks in Test Chat is what sold me — I could actually debug a bad answer instead of guessing.”
Marcus T.
Developer relations
“Swapping in our own Anthropic key took thirty seconds and the answer quality jumped immediately.”
Elena V.
Support lead
Start free, upgrade when you outgrow it
No card required. Bot, document, and message limits are enforced clearly at the point of use — never a silent failure.
Compare all plans- 1 bot with up to 20 documents
- 100 widget messages / month
- Default Ollama-backed model, no key required
- Full Test Chat + retrieval debugging