Turn your agent into a
go-to-market engineer.
Coil gives Claude, ChatGPT, or Cursor everything it needs to build and run outreach campaigns: ICP lists, signal-based sequences, sends through SmartLead or Instantly, and reports it actually acts on. You write the brief. It ships the campaign.
A recorded campaign, replayed verbatim — brief to booked calls. Sign up and brief your own agent in about 90 seconds.
It doesn’t assist with campaigns.
It owns them.
Fourteen skills covering the whole go-to-market loop — scope, build, launch, measure, iterate. You set the goal and approve the big calls. Your agent carries the pipeline.
Scopes the campaign
Give it a launch, a market, and a goal. It compiles the ICP, pulls a ranked list with live funding and hiring signals, verifies every email, and proposes the plan.
Engineers the sequence
Openers built on what actually happened — the round, the hire, the launch — in your tone, with A/B variants per lead so the campaign learns as it runs.
Runs it & iterates
Launches through your SmartLead or Instantly account, triages replies, reports against the goal — and shifts volume to whatever is winning, without being asked.
Five primitives.
Infinite plays.
Scraper, enricher, verifier, sequencer — and the AI driving them. Coil’s skills are building blocks, not a fixed pipeline. Describe an outreach motion nobody’s productized yet, and your agent chains the tools together itself.
The first-sales-hire play
Find SaaS companies that just posted their first AE role — that founder is about to care a lot about pipeline.
The unhappy-rival play
Watch review sites for frustrated customers of our competitor and reach out within 24 hours.
The conference play
Everyone speaking at European fintech events this autumn — opener references their actual talk.
Everything the agent can do,
you can do. And vice versa.
Coil isn’t a chatbot bolted onto a database. Every capability ships three ways — dashboard, CLI, and agent skill — backed by the same API. Switch mid-task without losing anything.
A real dashboard, not a chat log
Filterable lead tables, sequence editors, automation rules, reply inbox. When you’d rather click than type, everything is right there.
sequences 3 running · 62% open
replies 7 awaiting triage
Skills, not screenshots
MCP skills your AI discovers and calls natively — same auth, same org, same data. It’s a teammate with its own keys, not a macro clicking buttons.
coil.signals.fetch(company_ids)
coil.sequence.queue(list, provider) ✓
Built on org-scoped API keys with audit logging — your agent acts as a member of your team, with the same permissions model as everyone else.
The three things it replaces
in your stack.
We’re honest about what each one’s good at — and what still lands back on your plate.
A database, not an engineer.
Apollo gives you contacts. It doesn't scope a campaign, write openers, or tell you what's working. You're still the person who operates it on Sunday night.
→ Coil runs the whole loop end-to-end, from one chat or one dashboard.
Powerful, if you learn it.
Clay rewards people who can build workflows. It's the right tool if you have a RevOps engineer. If you're the founder doing outreach yourself, it's another job on top of your job.
→ With Coil, the workflow engineer is your agent — it already knows how to chain the tools.
$3k/mo, slow, inconsistent.
A contractor takes a month to ramp, costs $2-5k/mo, gets sick, takes holidays, and forgets follow-ups. The reports are mostly vibes. You suspect the reply rates are made up.
→ Coil is ready Monday and never forgets a follow-up. You drive it, or your agent does. Same platform.
Questions founders actually ask.
If something else matters more, ping us. These are the honest answers, not the FAQ-ish ones.
Do I need to use an AI agent to use Coil?
No. The Coil dashboard does everything skills can: list building, sequence sending, automations, reporting. Agents just make it faster because they handle research, personalization, and reply triage in chat instead of clicks. Use whichever fits the task; the data and actions are the same.
Will my emails sound like AI wrote them?
They sound like you. Coil learns your past replies and matches tone, whether you draft in the dashboard or hand it to an agent. The openers reference real signals (a Series A, a new hire, a product launch), not "I noticed your impressive background." If a draft sounds generic, rewrite it in your voice, either in the editor or in chat.
Whose AI bill am I paying, yours or theirs?
If you drive Coil through an AI: theirs. Coil ships as MCP skills your existing AI calls. You're already paying for Claude or ChatGPT, and Coil makes that subscription do more. If you'd rather skip AI and just use the dashboard or CLI: no AI bill at all. We charge separately for the lead-data plumbing either way, and it's free during beta.
What about deliverability?
Sends run through your SmartLead or Instantly account. Coil doesn't touch your sending domain. We hand verified, deduped leads to the tool you already use for warming and inbox placement. The agent reports inbox-eligibility before queueing.
How fresh is the lead data?
Sales Nav scrapes run live. Funding and hiring signals refresh daily. Coil tells you when a record's older than two weeks so you can decide what to do with it: re-verify, skip, or send anyway.
What happens when beta ends?
30 days' notice and a public usage rationale before any billing change. No card collected today means no surprise charge later. If pricing doesn't fit you when it arrives, you walk, and we'd rather know why than keep you for one more cycle.
Your GTM engineer
starts today.
No recruiter, no ramp-up, no salary line. Sign up, write the brief — or drive the dashboard yourself — and launch your first campaign this afternoon.