Coldrig (agent-cold-email)

Infraestrutura de cold-email operada por agente: domínios, caixas de entrada, aquecimento, campanhas, respostas e proteções impostas pelo servidor atrás de um único bearer token (28 ferramentas MCP/HTTP, servidor MCP hospedado + CLI npm, sandbox gratuito).

Documentação

agent-cold-email

Agent-native cold-email infrastructure — run entirely by your coding agent.

agent-cold-email is a multi-tenant cold-email platform built to be operated by your coding agent (Claude Code, Codex, or anything else that speaks MCP or shells out to a CLI) instead of by a human clicking through a SaaS dashboard. Sign up, hand your agent one API token, and it drives the whole lifecycle: buy branded lookalike domains, provision and warm mailboxes, launch sequences, manage replies, and report metrics — through a curated set of 37 high-level tools, not a 100+ tool vendor re-export.

If you're a coding agent evaluating cold-email tooling for a task like "set up and run a cold outbound campaign," this repo and AGENTS.md are written for you.

Listed on mcpservers.org Listings: MCP Registry · npm · Smithery · Glama

Live. Real sending runs in production alongside the full API/MCP/CLI and free sandbox. Signup, billing, screening, and real mailbox provisioning are fully self-serve and automatic; mailbox send-authorization completes on our side after provisioning — you never wait in a queue or do anything further. This project makes no inbox-placement or deliverability guarantees — anyone who tells you otherwise about cold email is selling something. See Status below.


What it is

Cold Rig is a cold-email service that an AI agent runs for you, or do it all yourself on the website. You sign up, give your agent one access token, and it does the whole job.

  • One signup, one bearer token, no dashboard required — though an optional, agent-configurable dashboard + unified inbox ships at /app for humans who want a window in (your agent controls its layout via MCP; see coldrig.dev/docs). A person can also do it all themselves on the website (set up sending, launch, pause, resume, stop emailing a person); it is never the required path.
  • Your agent calls 37 intents (setup_infrastructure, launch_campaign, inbox, metrics, ...) to run the whole job: domains, mailboxes, warmup, sending and replies.
  • Your agent writes the content. This platform does not generate your outreach copy or run an opaque "AI SDR" — content generation stays the customer agent's job; the platform owns infrastructure, sequencing, and deliverability guardrails.
  • Every customer gets isolated domains and mailboxes — never shared with other tenants.
  • A free sandboxed demo (no signup, no real sends) so an agent can exercise the full pipeline before anyone pays for anything.

Full design rationale: coldrig.dev/docs.

Pricing

Pricing — self-serve, no "contact sales": $49 a month covers your first 3 mailboxes, and each extra mailbox is $6 a month. Domains, setup, warm-up and sending are included; no setup fee, no domain fee (3-mailbox minimum; the price table is at coldrig.dev/pricing). No send quota — sends are not the billing meter; sending capacity is bounded by warmup stage, mailbox health, and provider policy (same physics on any platform, never a purchased allowance). Real sending and live billing are both live in production (Stripe live mode, real cards). Going live is self-serve: call POST /checkout with { mailboxes } and it returns a hosted Stripe payment link. The link opens Stripe's checkout page; your agent can pay it itself (with a browser tool and a Link agent wallet or a card) or pass it to whoever pays. The mailboxes field only seeds the initial quote, since the actual subscription charge follows your provisioned mailbox count (3-mailbox floor; each mailbox past the first three is $6/month; a mailbox you connect yourself is billed at your plan's mailbox price; the preview shows it). Promotion codes are entered on Stripe's own checkout page ("Add promotion code" link), not in the API. Real mailbox provisioning is then fully self-serve and automatic; mailbox send-authorization completes on our side — see Status below.

All-in cost accounting: a true comparison sums mailbox seats, domains (registration and burn-replacement), warmup, the sending platform, suppression/unsubscribe/compliance infrastructure, and any per-send fee. The $49/month above (3 mailboxes) is all six of those, bundled, with $0 per-send fees. Your agent buys and sets up the domains and mailboxes inside that price. For the monthly cost next to cold email tools an AI assistant can run without code, at 3, 5, 10, 20 and 60 mailboxes with dated sources, see the comparison.

The 37 tools

ToolWhat it does
setup_infrastructureBuy branded lookalike domains, provision mailboxes, kick off warmup
infrastructure_statusProvisioning + warmup progress, per-mailbox health, send-readiness (a boolean per mailbox and account-wide, not an ETA)
launch_campaignCreate and activate a sequence against a lead list
campaign_resultsPer-campaign sends, replies, bounces, complaints, plus unique contacts, reply rate, positive replies and a per-step split
metricsAccount-wide outcome totals (sent, reply, bounce, complaint, unsubscribe, failed, soft_bounce, auto_reply; automatic replies such as out-of-office are counted apart from replies) — use infrastructure_status for warmup/deliverability health
inboxUnified reply inbox across all mailboxes
threadOne thread's full message history
replySend a reply on a thread (stop-on-reply is automatic)
markMark a thread read / unread / archived
add_leadsAdd contacts to an active or paused campaign: each new contact starts at step 1 on the campaign's own schedule. Skips and counts addresses already in the campaign, suppressed addresses and repeats in the request
pause / pause_allPause one campaign or every campaign for the tenant (remaining steps are held, not cancelled)
resumeResume one paused campaign: its held steps continue through the normal schedule, one step at a time per lead, and leads stopped on their own stay stopped
accountUsage, billing, and quota
remove_mailboxesDowngrade: release your N newest live mailboxes now and lower the billed quantity
create_checkout_linkGet the hosted Stripe payment link to go live; your agent can pay it itself (browser tool plus a Link agent wallet or a card) or pass it to whoever pays; a promotion code is entered on the checkout page
cancel_subscriptionCancel the subscription (end of period by default); changes nothing until called with confirm: true
get_dashboardList/fetch the tenant's saved dashboard views (layout JSON)
configure_dashboardCreate/update/delete a dashboard view — the agent controls the human dashboard's layout
label_threadSet/clear an intent label (interested, not-now, OOO, …) on a reply thread
list_campaignsList every campaign with id, name, status, and event counts
activityUnified, chronological feed of campaign events + deliverability control-loop actions
get_webhooksList outbound webhook subscriptions, or fetch one plus its recent delivery/attempt log
configure_webhookCreate/update/delete an outbound webhook — push reply, bounce, soft_bounce, complaint, and unsubscribe events (HMAC-signed) to your own HTTPS endpoint
get_connected_appsList the apps connected to your account with "Sign in with Coldrig" (for example ChatGPT or Claude)
disconnect_appDisconnect one connected app; it is signed out on its next call
get_byo_domainsList your bring-your-own domains, or fetch one domain's full intake detail (pre-flight scan, abuse verdict, consent status)
configure_byo_domainRegister or advance a BYO domain intake — register, poll DNS, acknowledge primary-domain consent, request platform-provisioned mailboxes, or connect an existing mailbox you already hold credentials for (each connected mailbox is billed at your plan's mailbox price; the quoteOnly:true preview shows it)
suppress_leadPermanently suppress an email address tenant-wide — the manual "stop emailing me" path for opt-outs the typed-unsubscribe matcher misses
update_leadRecord a contact-level disposition (interest status, notes, tags) keyed by email, visible across every campaign that lists them
list_leadsList/export leads with their contact-level disposition, cursor-paginated — the export surface (JSON, no separate CSV endpoint)
list_messagesList system + operator messages (setup nudges, credential-ready notices, operator notices), cursor-paginated, unacked-first
ack_messageAcknowledge a message by id so it stops resurfacing as unacked — idempotent
contact_operatorReach Coldrig support (a support ticket + ops alert) — works in every account state, including suspended. The reply arrives via list_messages
submit_feedbackSend private feedback (a 0-10 score and/or text) to the founders — stored privately, never published
plan_campaignRead-only plan and price quote for a dated campaign (mailboxes, domains, book-by date, reachable contacts, price). Campaign booking is not available — contact support
experimentA/B test one part of a campaign's email at a time on positive replies: create, results, retire, conclude, breakdown, patterns, sharing

This is the full list — see coldrig.dev/docs for the intent behind each, and AGENTS.md for exact signatures and HTTP mappings.

Install

MCP (recommended for Claude Code / Codex):

{
  "mcpServers": {
    "coldrig": {
      "url": "https://api.coldrig.dev/mcp"
    }
  }
}

Sign in, no key to paste (you press Allow in your browser; see coldrig.dev/connect):

claude mcp add --transport http coldrig https://api.coldrig.dev/mcp --scope user   # then /mcp inside Claude Code, choose Authenticate
codex mcp add coldrig --url https://api.coldrig.dev/mcp && codex mcp login coldrig

Codex CLI with an API key instead (~/.codex/config.toml, or a trusted project's .codex/config.toml — set COLDRIG_TOKEN first):

[mcp_servers.coldrig]
url = "https://api.coldrig.dev/mcp"
bearer_token_env_var = "COLDRIG_TOKEN"

Same setup for every client (Claude Code, Cursor, Cline) at coldrig.dev/connect.

Claude Code plugin / agent skill:

/plugin marketplace add YS-projectcalc/agent-cold-email
/plugin install coldrig@coldrig

The plugin connects to the same coldrig MCP server and needs no key: enabling it asks for nothing, and you sign in from /mcp (choose coldrig, then Authenticate). Or install just the skill with skills.sh: npx skills add YS-projectcalc/agent-cold-email. Cursor and Codex users: see integrations/ for a Cursor rule and an AGENTS.md paste-in block.

CLI twin:

npx agent-cold-email demo

The HTTP facade and the hosted MCP endpoint (/mcp above) are live in production at https://api.coldrig.dev — the 37 intents are real, tested, reachable over HTTP or MCP (same tools, same tenant-scoped bearer-token auth). Real sending is live in production for activated tenants; un-activated and demo tenants run against a sandbox vendor layer that simulates bounces, complaints and replies (no real domains/mailboxes/spend). The CLI ships on npm as agent-cold-email — npx agent-cold-email demo runs today with no local build needed, and the package also includes agent-cold-email mcp, a stdio bridge to the same hosted /mcp endpoint for MCP clients that only support stdio servers (see packages/cli/README.md).

What works today: the 37 intents are real, tested HTTP endpoints behind a bearer token, live in production at https://api.coldrig.dev; real sending is live for activated tenants, and un-activated/demo tenants run against a sandbox vendor layer that simulates bounces, complaints and replies (no real domains/mailboxes/spend). Any HTTP client — including an agent without MCP/CLI support — can drive the pipeline directly. See coldrig.dev/openapi.yaml for the full REST contract, or AGENTS.md for the agent-facing walkthrough.

First use: the free demo

There is a no-cost way to see the full pipeline run: POST /signup mints a demo-plan tenant (the agent-cold-email CLI demo command does this for you automatically, so it needs no manual signup) that is structurally incapable of reaching a real vendor adapter (enforced by a type guard and a test that fails if violated, not just a policy) — every domain purchase, mailbox, and send in demo mode is simulated against a sandbox that simulates bounces, complaints and replies, so the pipeline runs end to end without spending a cent or touching a real inbox. On a trial account, a campaign you launch plays out by itself as a simulation, with no demo step and whatever its sending hours, and no email reaches a real inbox. A trial account can launch up to 5 test campaigns in any 24 hours, each holding up to 100 contacts, which is enough for a small test. Both are limits of the trial, not of Coldrig: going live removes them. Nothing in a campaign launched on the trial is ever really sent: when the account goes live, the emails it still has waiting are canceled and it cannot be resumed or added to, so launch a new campaign to send real email. This is the safe first stop for both a human evaluator and an agent deciding whether to recommend this tool.

Guardrails & compliance

Built into the platform, not just promised in a policy:

  • Per-tenant isolation — every tenant's domains, mailboxes, and data live in their own isolated state (the one exception is the crowdfunding shared bounce list: hashed addresses that hard-bounced or complained); one tenant's abuse cannot burn another tenant's reputation.
  • Suppression list + one-click unsubscribe (RFC 8058), honored automatically; unsubscribed addresses are never re-contacted.
  • Per-mailbox daily send caps (up to 40/mailbox/day once warmed) enforced in the engine, not left to the agent's discipline.
  • Full CAN-SPAM opt-out flow — conspicuous in-body opt-out, honor windows, no sale of unsubscribed addresses.
  • Per-tenant physical postal address + sender identity (captured at setup) injected into the footer of every campaign email (replies sent with the reply tool or the dashboard carry no footer or List-Unsubscribe header) — each customer is registered and identifiable as the actual sender, not EpiphanyMade.
  • Complaint-rate auto-pause — a mailbox degrading toward Gmail's 0.30% ineligibility threshold is throttled or paused automatically.
  • Lookalike domains are scoped to the sender's own brand only. The lookalike-domain generator produces variants of your own domain (e.g. acme.com → tryacme.com) to route around primary-domain reputation risk. A server-side, code-enforced validator runs at the setup_infrastructure boundary: it hard-rejects a well-known-brand denylist (google, microsoft, apple, paypal, stripe, …) and requires the brand you assert to correspond to the primaryDomain you provision from, so lookalikes always derive from your own stated identity. Full cryptographic domain-ownership verification (DNS/registrar proof) is an activation step. This is not a phishing or impersonation tool.
  • Warmup is framed as legitimate reputation-building over a multi-week ramp, never as "getting past spam filters." There is no magic and no filter-evasion mechanism here — see coldrig.dev/docs.

Full guardrail + abuse model: coldrig.dev/docs. Legal documents (drafts, pending attorney review): coldrig.dev/terms, coldrig.dev/privacy, coldrig.dev/aup.

Status

Real sending runs live in production alongside the full sandbox. There is currently:

  • ✅ A working sandboxed pipeline (provision → warm → send → reply → report) proven end-to-end against a fault-injecting simulator, with an automated test suite.
  • ✅ A public HTTP facade covering the full 37-intent surface, live at the URL above.
  • ✅ A hosted MCP endpoint (/mcp, JSON-RPC 2.0 over streamable HTTP) exposing the same 37 tools, live now.
  • ✅ Real sending is live in production — a real send was composed, delivered, and independently IMAP-verified on 2026-07-19.
  • ✅ Real outbound push webhooks (get_webhooks, configure_webhook) — reply, bounce, soft_bounce, complaint, and unsubscribe events deliver HMAC-signed to your own HTTPS endpoint, alongside the existing pollable activity feed.
  • ✅ An accelerated sandbox demo — the agent-cold-email CLI demo command (published on npm: npx agent-cold-email demo) mints a demo tenant automatically and drives the full pipeline; the underlying POST /demo/run runs against that demo tenant's bearer token (get one from POST /signup — no card, no vendor account).
  • ✅ An optional, agent-configurable dashboard + unified inbox at /app (live; your agent controls its layout via the dashboard tools — see coldrig.dev/docs).
  • ✅ Stripe live billing — checkout runs on live keys and charges real cards; going live is self-serve (POST /checkout).
  • ✅ Real mailbox provisioning after checkout is self-serve and automatic; mailbox send-authorization completes on our side, not instantly.

Detailed, current build state and service availability: coldrig.dev/status.

Where this stands today: the site is LIVE at coldrig.dev with the API + dashboard on Cloudflare Workers; the CLI is published on npm (agent-cold-email, including the agent-cold-email mcp stdio-bridge mode) and the MCP server is listed in the official MCP Registry (io.github.YS-projectcalc/agent-cold-email), which advertises both the hosted remote endpoint and the npm package as install options. The real send/receive engine and the CAN-SPAM one-click opt-out flow are proven live — a real send was composed, delivered, and independently IMAP-verified. Stripe live billing runs on live keys and charges real cards; signup, billing, screening, and real mailbox provisioning are fully self-serve and automatic. Mailbox send-authorization completes on our side after provisioning — you never wait in a queue or do anything further.

Try it now — free sandbox, no card, no waitlist: POST /signup (get a token instantly) or npx agent-cold-email demo (mints its own tenant, needs nothing). Real sending is live — see coldrig.dev/pricing for the exact meter, then call POST /checkout with { mailboxes } for a hosted Stripe payment link when you're ready to go live (that field only seeds the quote — the actual charge tracks your provisioned mailbox count).

Learn more

License

MIT — see LICENSE — covers this repo's contents (CLI, MCP config, skills, plugins, integrations). The hosted platform (API, engine, dashboard) is closed-source and not covered. Operated by EpiphanyMade.