CloseHunt is a multi-channel AI sales platform — its agents find buyers showing real intent, score them against your ICP, then run outreach, AI auto-reply and 24/7 inbound listening across 8 channels. GDPR-clean.
Replace your SDR stack with agents you configure end-to-end. Each one with its own scope, goal and channels. Run as many as you need, in parallel.
Each agent watches one source forever · funding rounds, hiring spikes, tech-stack changes, news, LinkedIn engagement. No prompts, no babysitting, no missed signal.
Reply on email Monday, on WhatsApp Tuesday. The AI keeps one continuous thread across every channel a prospect uses · with a brand-aligned identity, goal-driven replies and a memory that updates on its own.
The CloseHunt score blends how well a lead matches your ICP with their actual engagement · market signals, replies, clicks, opens, intent picked up by the AI. It moves in real time and you always see the per-criterion breakdown (industries 96%, size 88%, geo 35%…). No black box, no opaque number.
Discovery → enrichment → qualification → outreach → reply → booked. Every stage is a live database state, streamed to your dashboard via SSE within seconds. No refresh button, ever.
Mount an agent on any inbox · email, LinkedIn DMs, WhatsApp, SMS · and it answers under your brand identity. It auto-creates the lead, patches their profile from every message, yields if another agent is actively closing, and quietly rewrites its own playbook from every outcome.
Each agent watches its own outcomes · which signals converted, which intros worked, which objections killed deals · then rewrites its own playbook every 6 hours. One for search, one for replies. Per-agent isolated, kept when you duplicate.
Replies, clicks, opens, call bookings, unsubscribes. Every interaction is logged, per lead, per channel.
Every 6 hours, your agent rereads its recent results and pulls out two clear summaries · the angles, signals and openers that actually convert for you, and the ones to drop.
From the next campaign on, your agent prioritises the right signals, opens with sharper hooks and avoids the angles that didn't land. You see performance climb without touching the prompt.
Beyond rigid sequences, your agents can plan the next touch lead by lead · channel, timing and angle chosen from live context, bounded by your caps. No sequence to maintain.
One agent picks the right surface per prospect, no duplicates, no spam, with cross-channel state baked into every send.
Bring your own API keys. 13 enrichment providers (Apollo, Hunter, Clearbit, Lusha, ZoomInfo, and more), plus HubSpot, Slack and Stripe. One sign-in, one bill, one source of truth.
A live walkthrough. The agent navigates, scores, sends, books. You step in when the deal is yours to close.
Every signal pulls from public news, sanctioned APIs or your own webhooks. LinkedIn outreach ships as manual-assist drafts by default. Flip on Autopilot, capped, audited, with explicit risk acknowledgement.
No public pricing for now. Talk to our sales team and we'll size a trial workspace to your volume, your channels and your use case. Everything below is included from day one.
CloseHunt is a multi-channel AI sales platform. You configure agents in one of four modes · SEARCH (discover new leads), OUTREACH (contact existing leads), BOTH (full lifecycle), or LISTEN (claim an inbox and reply 24/7 under your brand voice). They score against your ICP, run personalised outreach across 8 channels, and auto-reply until the goal you set is reached.
Eight: Email (your connected mailbox), LinkedIn (manual-assist or Autopilot), WhatsApp, Telegram, Instagram, Messenger, SMS and outbound AI voice calls. Every channel runs through the same router with caps and cross-channel state.
Yes, by design. Every prospect comes from public APIs or sanctioned providers. The platform ships with consent log, public opt-out, hard-bounce auto-opt-out, data-subject endpoints, retention windows, and audit log on every action.
No. LinkedIn outreach is manual-assist by default. AI writes drafts, you click 'Open LinkedIn' to send. Autopilot is opt-in with a safety layer: warm-up ramp, per-account caps, trust score, cool-off after incidents, business-hour gating.
Every inbound message is classified by intent (positive / negative / call request / closing request / unsubscribe / out-of-office). The AI generates a goal-driven reply through the same channel. Goal modes are tri-state · pick a preset (book call, qualify, close), write your own free-text goal, or run unbounded for support-style conversations. The agent pauses itself when the goal is reached, never before.
Talk to our sales team and we'll size a trial workspace to your volume, your channels and your use case. No public pricing for now, everything is calibrated to your needs.
An agent that doesn't hunt · it sits on one or more inboxes you assign it (email, LinkedIn, WhatsApp, SMS, you pick) and answers every inbound under its identity. It auto-creates a lead when the sender is unknown, patches the lead profile from every message, and politely yields if another agent is actively driving the conversation. Perfect for support, post-sales, or as a brand-aligned receptionist for your social inboxes.
Opt-in per agent. The agent reads its own outcomes · which signals converted, which intros worked, which objections killed deals · and rewrites two short playbooks every 6 hours: one for search, one for replies. Cheapest model, hot-path safe, per-agent isolated, and the insights survive duplication. Your agent quietly gets better at its job without you touching the prompt.
Yes. Agent Identity carries name, title, company description, mission, operating directives, escalation contact, and context notes · all bounded so the prompt budget never blows up. Paste your website URL and the AI extracts a first draft for you. One identity can be reused across many agents, so your whole fleet sounds coherent. Directives are binding in the reply prompt, so the agent never goes off-script.
Yes. A single SSE bus pushes 13 typed event streams · new lead, hot lead, reply received, message sent, agent run finished, goal reached, signal captured, and more · directly to every open tab. Notifications, the unibox, the prospect timeline, agent panels and pipeline counters all refresh within a few seconds. No refresh button, no polling cost on your side.
Your sequences run in two modes, chosen per sequence. In AI mode (default), every message is drafted on the fly by the AI for that specific prospect · highest quality, ideal for audiences that need real personalisation. In Template mode, you write the body once with variables and the engine just does substitution · zero AI calls on first send, ideal for high-volume campaigns. Inbound replies always go through the AI regardless of mode. You can switch any time from the sequence settings.
Yes. CloseHunt is exposed as a tool to Claude Desktop and any MCP-compatible client. Your AI tools can search leads, draft messages, enrol prospects in a sequence, drop a signal, list your sequences and pull analytics · workspace-API-key auth, full audit log on the CloseHunt side.
It never makes things up. When the prospect asks something out of scope (a specific price, a contract clause, a deep technical question), the agent pauses its auto-reply on that lead. It alerts your rep via Slack, email and in-app notification with full context (the last 5 turns, the detected intent, and a direct link to the conversation). Your rep answers in their own voice from their own inbox or from the app. When the prospect replies next, the agent picks the thread back up, aligning to the tone and angle the rep just set. No lost thread, no invented reply.
Yes, on three levels. (1) The agent's identity carries strict directives the AI can't bypass · "never discuss pricing without checking with me", "escalate any security question". (2) You choose what happens on negative replies · immediate stop on the first firm no, or one more attempt with a negative-streak cap you set. (3) You can also force a hand-off to your rep after a fixed number of turns, even if the goal isn't reached yet. All settings are per agent.
Trial on request · contact our sales team. Deploy your first agent in under 3 minutes.