AI responses grounded in your own docs. Unlimited agents at a flat price. Every channel in one inbox. Built for teams who need it to just work.
Scale from 2 to 200 agents at the same flat price. Onboard seasonal staff, bring contractors in for peak periods, or add your whole leadership team without a budget conversation.
One plan
$20
per month
Unlimited agents
Unlimited AI
All channels
Campfire's AI is grounded in your knowledge base. It never invents answers. Track your own resolution rate in the dashboard, so the only number you trust is the one your team sees.
Connect OpenAI, Anthropic, or any OpenRouter-compatible model. You hold the keys — we never proxy your data through a shared endpoint.
AI answers come from your actual docs, not hallucinated knowledge. Semantic search retrieves the right chunks before every reply.
Routine support questions get resolved by AI before reaching the queue. Your dashboard shows the per-workspace rate so you can verify it instead of taking our word for it.
Compose-time suggestions surface as your agents type — not as a separate AI tab. One click to accept, zero friction.
// Point at any OpenRouter-compatible model
export const aiConfig = {
provider: 'openai',
model: 'gpt-4o-mini',
apiKey: process.env.OPENAI_API_KEY,
ragEnabled: true, // grounded in your KB
}Web chat, WhatsApp, Telegram, email — all in a single shared queue. No per-channel add-ons, no separate dashboards to monitor.
One script tag. Fully customisable. Works with React, plain HTML, or any SPA — no wrapper library required.
Incoming messages from WhatsApp Business and Telegram land in the same inbox thread as your web conversations.
Forward your support address and every reply thread lives alongside real-time chat. One team, one queue.
No app switching. Every channel surfaces as a conversation row. Filters, assignment, and tagging work the same regardless of origin.
Write articles once. They surface in your help center, in the widget, and as the retrieval layer behind every AI response — no separate indexing step.
Write once, publish everywhere. Articles are indexed for semantic search and embedded directly in the widget help panel.
Keyword search misses intent. Our search understands meaning, so customers find the right article even when the words don't match.
Before a visitor sends a message, the widget surfaces relevant articles automatically. Most questions answer themselves.
Organise content into collections and categories. Drag-and-drop ordering. SEO-friendly URLs out of the box.
Resolution rates, CSAT, agent performance, response times, and AI usage — all tracked automatically. No event instrumentation required.
See exactly how many conversations were resolved by AI, by agent, or escalated — broken down by channel and time range.
Post-conversation satisfaction surveys are on by default. Scores aggregate per agent, per team, and per product area.
First response time, full resolution time, and time-in-queue are tracked automatically. No manual tagging required.
Individual and team-level metrics: volume handled, CSAT, resolution rate, and average handle time — all in one view.
REST endpoints, webhooks, typed tRPC procedures, and an embed script you can drop into any page with a single line.
One-line embed
Drop a single script tag into your HTML. No npm install, no bundler config.
REST + webhooks
Full REST API with webhook delivery for every conversation event. Signed payloads, automatic retries.
SDK packages
TypeScript-first client packages for Node and browser. Auto-generated from the tRPC schema.
API key management
Per-workspace API keys with scoped permissions. Rotate without downtime.
<!-- One line. That's it. -->
<script
src="https://cdn.campfire.is/widget.js"
data-workspace="your-workspace-id"
async
></script>We will not claim audits we have not finished. Here is the honest snapshot of where we are — encryption in transit on every workspace, RLS-enforced tenancy, and a SOC 2 Type II audit underway.
SOC 2 Type II — underway
Audit in progress. Status available from your account contact on request.
GDPR-friendly
Data export and deletion endpoints on request. DPA signable for EU customers.
Tenant isolation by RLS
Multi-tenant data isolated at the Postgres layer using Supabase row-level security on every query.
Encryption in transit
All traffic between widget, dashboard, and backend is encrypted with TLS 1.2+.
Row-level security on every database query. Multi-tenant isolation enforced at the Postgres layer — no shared data, no cross-customer leakage.
Intercom charges $85 per seat, then adds $0.99 for every AI resolution. A 10-person team handling 1,000 AI conversations pays over $1,000/month before any add-ons.
Start your free 14-day trial today. No credit card required. Unlimited seats from day one.