HeyLead
HeyLead is an MCP server for AI-driven LinkedIn outreach, campaign management, and inbox/reply operations.
Set up HeyLead, connect LinkedIn, and build a voice signature for personalized outreach.
Generate ICPs/personas for sales, hiring, job search, partnerships, vendors, or research.
Create and edit campaigns with target descriptions, saved ICPs, named people, project briefs, language, and sending settings.
Launch, pause, resume, archive, delete, emergency-stop, retry, or repair campaigns.
Find, view, skip, close, or dismiss prospects and track conversations/outcomes.
Send follow-ups, replies, InMails, and inbox answers; approve or discard drafted replies.
Check replies, classify sentiment, detect opt-outs, and surface hot leads.
Read LinkedIn inbox conversations, drafted comments, and files people send.
Search, browse, export, tag, note, re-stage, enrich, or link contacts.
Book Google Calendar meetings with prospects and send invites.
View dashboards, campaign progress/history/plans, analytics, comparisons, and exports.
Manage scheduler status/settings, cloud/local sending, reports, logs, and diagnostics.
List, switch, unlink accounts, connect email accounts, and refresh LinkedIn tier.
Get next-action recommendations and view outreach status.
Allows sending outreach emails through a connected Gmail mailbox via Unipile.
Allows booking meetings by placing agreed calls on Google Calendar and sending the prospect an invite.
Syncs campaign contacts and deals to HubSpot CRM.
This server has a remotely hosted endpoint. Open its connector listing for connection details, authentication requirements, and setup instructions:
HeyLead
HeyLead is an AI agent for LinkedIn outreach: it finds the right people, writes to them in the voice of your own LinkedIn posts, follows up, and handles replies. It runs from Claude Code, Cursor, any MCP client or a web dashboard. Connect: https://heylead.dev/mcp (OAuth; works in Claude Code, in claude.ai as a custom connector, in Cursor, and in ChatGPT as a custom MCP app — tested on ChatGPT Pro on 29 September 2026; custom MCP apps depend on your ChatGPT plan and settings). Website: heylead.dev · Agent setup and docs: heylead.dev/llms.txt
If HeyLead finds people for you, star the repo so the next founder can find it.
Use cases
A campaign has one of six goals, and the ICP, the fit check and the messages follow it: Sell a product or service, Find a job, Hire people, Find partners or investors, Find a vendor and Research interviews.
Sell a product or service: Reach the people who buy what you built.
Find a job: Reach the people who hire for the role you want.
Hire people: Reach candidates for a role you are filling. Works with a custom brief.
Find partners or investors: Reach the people who can sign a partnership or an investment. Works with a custom brief.
Find a vendor: You are the buyer. Reach the people who sell what you need.
Research interviews: Reach people to interview, survey or test with. Works with a custom brief.
Hire people, Find partners or investors and Research interviews run on a custom brief until their message sets exist.
Related MCP server: LinkedNav MCP
Getting Started
MCP (Model Context Protocol) lets AI assistants use external tools. HeyLead gives your AI the ability to do LinkedIn outreach for you.
You need: Claude Code, claude.ai, ChatGPT, Cursor, or any MCP client — or use the web dashboard at heylead.dev/dashboard.
Step 1: Install HeyLead
Hosted remote server (nothing to install): paste Set up HeyLead using https://heylead.dev/llms.txt into your AI chat, or add https://heylead.dev/mcp in Claude Code:
claude mcp add --transport http heylead https://heylead.dev/mcpClaude Code, as a plugin (hosted, nothing to install):
/plugin marketplace add D4umak/linkedin-outreach-mcp
/plugin install heylead@heyleadThen run /mcp, pick plugin:heylead:heylead and choose Authenticate to sign in to your
HeyLead account. Use only one HeyLead connection, not both: if you also added HeyLead as a connector on claude.ai, sign that one in and skip the plugin's server, or remove it. Or sign in without the menu:
claude mcp login plugin:heylead:heyleadThe plugin connects the hosted server at
https://heylead.dev/mcp and adds skills for starting a campaign, approving
waiting messages, answering replies, a weekly review, pausing, and adding a
list of people. Its source is in
plugins/heylead.
Or run the client locally over stdio. Install the CLI once with uv:
uv tool install heyleadClaude Code:
claude mcp add heylead -- heyleadCursor: Settings > MCP > "Add new MCP server" > Name: heylead, Command: heylead
Any MCP client:
{
"heylead": {
"command": "heylead",
"args": []
}
}Update with uv tool upgrade heylead.
Step 2: Set up your account
HeyLead connects to heylead.dev. Take the 90-second quiz or sign in at heylead.dev/dashboard/login. Your quiz personas become a draft campaign. Connect LinkedIn in Settings → Connected accounts, then launch from the dashboard — or copy your setup message from Settings → Integrations → Chat client into your AI chat. Nothing sends before you launch. Hosted users share a professional directory; campaigns and inboxes stay private.
Step 3: Find leads
"Find me CTOs at fintech startups in New York"
"Send outreach to the campaign"
"Check my replies"
"How's my outreach doing?"How It Works
Define your ICP — "Generate an ICP for AI SaaS founders" → RAG-powered personas with pain points, barriers, and LinkedIn targeting
Create a campaign — "Find me fintech CTOs" → searches LinkedIn, scores prospects by fit
Warm up prospects — Engages with their posts (comments, likes) before reaching out
Send personalized invitations — Voice-matched messages that sound like you, not a bot
Follow up automatically — Multi-touch sequences after connections are accepted
Handle replies — Detects sentiment, advances positive leads toward meetings, answers questions
Track outcomes — Won/lost/opted-out tracking with conversion analytics
Sending model: campaigns are created as drafts and only start when you
explicitly launch them. Every send passes rate limits, working-hours checks,
and a 1st-degree connection guard before it goes out. HeyLead sends from your own LinkedIn account at a human pace: at most 20 invitations a day and 100 a week on a free LinkedIn account (more on Premium or Sales Navigator: a newly connected paid account starts at 20 a day and doubles each week up to its limit), Monday to Friday 08:00 to 22:00 in your time zone, minutes apart. It backs off when LinkedIn pushes back and resumes on its own. You can pause any campaign at any time. Launching is also what
commissions cloud sending — in observe mode nothing is commissioned and
nothing is sent, from either machine.
Approval mode (hosted): a workspace that has not chosen otherwise sends
opening DMs, follow-ups and answers to replies as HeyLead writes them, without
anyone reading them first. Switch review on (Settings → Sending, or
scheduler(action="approval_mode", mode="require_approval")) and they wait for
a person, read on the Approvals page or with inspect(action="waiting"); the
approved text is what is sent. Invitations and InMail are never held. In review mode, invitations go out without a note. This is the same on every plan.
Tools
HeyLead gives your AI 22 tools:
Core Workflow
Tool | What it does |
| Connects LinkedIn and analyzes your writing style into a voice signature |
| Generates a rich Ideal Customer Profile with buyer personas |
| Previews which LinkedIn profiles a saved ICP matches, without creating a campaign |
| Compiles a targeting request (country ties, interests) into LinkedIn recall queries and profile-evidence scoring |
| Creates an outreach campaign (as a draft) from a natural language description |
| Generates a personalized LinkedIn message and sends it |
| Checks for new replies across campaigns, classifies sentiment, surfaces hot leads |
| Puts an agreed call on your Google Calendar and sends the prospect an invite |
| Your dashboard — campaigns, stats, hot leads, account health. Links to the matching heylead.dev/dashboard page and, on hosted accounts, attaches a snapshot card |
Outreach & Engagement
Tool | What it does |
| Sends follow-ups and replies to prospects |
| Sends email via a connected Unipile mailbox (Gmail/Outlook). Never Mail.app. |
| Comments on, reacts to, follows, or endorses a prospect to build trust |
| Browses and reads LinkedIn inbox messages directly |
| Processes unreplied inbox messages through the inbound pipeline |
| Generates and publishes a voice-matched post to LinkedIn, or publishes your approved words with a photo or a short video |
Campaign Management
Tool | What it does |
| Campaign lifecycle — launch, pause, resume, archive, delete, emergency stop, retry failed |
| Edits a campaign's name, mode, booking link, or context fields |
| Manages prospects — skip, close with outcome, view conversation or timeline |
| Imports prospects from CSV data into a campaign |
Insights & Analytics
Tool | What it does |
| Campaign analytics — reports, comparisons, and exports |
| Read-only digest of operator holds, strategist replans, closer decisions, reply skips, and gated jobs |
| Curates the knowledge base that grounds messages — lists, adds, removes, refreshes, and searches sources. Hosted only |
| Recommends the best next action, prioritized by impact |
| Views and analyzes buying signals — news, company engagement, website visits, profile viewers |
| Adds, removes, and lists signal keyword watchlists |
| Network intelligence — a reciprocal pool of members' connected accounts; join to use it |
Growth & Relationships
Tool | What it does |
| Analyzes and improves your LinkedIn personal brand |
| Views and restores LinkedIn profile change history |
| Tracks follow-ups with business partners, vendors, and investors |
| Searches, browses, and manages your global contact base |
| Syncs campaign contacts and deals to HubSpot CRM |
Automation & Account
Tool | What it does |
| Manages the autonomous scheduler — status, on/off |
| Local git checkout only — patch this repo and/or open a PR |
| Manages LinkedIn accounts — list, switch, or disconnect |
| Hosted orgs — list, switch, invite editor/viewer, create a client workspace |
Key Features
Voice Matching — Analyzes your LinkedIn profile and posts to capture your writing style. Every message sounds like you wrote it.
ICP Generation — RAG-powered pipeline that crawls company context, generates buyer personas with pain points, fears, barriers, and maps them to LinkedIn search parameters.
Autonomous Scheduler — Runs in the background, respects working hours and rate limits. On a hosted account, cloud is the default sender for every campaign. Launching commissions the cloud, so outreach continues with your laptop closed, inside the sending window: invitations, opening DMs, first-touch InMail, follow-ups, engagements, follows, endorsements, email fallbacks, prospect top-ups, auto-replies, inbound, warmup, signal collectors, and post-intel. Observe means nobody sends.
Engagement Warm-ups — Automatically engages with prospect posts before sending connection requests, building familiarity.
Pace — HeyLead sends from your own LinkedIn account at a human pace: at most 20 invitations a day and 100 a week on a free LinkedIn account (more on Premium or Sales Navigator: a newly connected paid account starts at 20 a day and doubles each week up to its limit), Monday to Friday 08:00 to 22:00 in your time zone, minutes apart. It backs off when LinkedIn pushes back and resumes on its own. You can pause any campaign at any time.
Outcome Tracking — Mark deals as won/lost, track conversion rates, identify stale leads, measure engagement ROI.
Pricing
Free: $0, one connected account (a LinkedIn seat or an email mailbox), cloud sending included, up to 2 follow-ups per person, until your 2nd hot lead. From then HeyLead keeps working the conversations already open and reaches new people on Pro.
Pro: $29 per connected account per month (a LinkedIn seat or an email mailbox). HeyLead keeps reaching new people after your 2nd hot lead, runs more than 1 active campaign and follows up to 5 times. Monthly, cancel any time, no setup fee.
A hot lead is a person whose reply HeyLead reads as interested.
A connected account is a LinkedIn seat or an email mailbox. Free includes one of either.
Invitation limits follow the LinkedIn account (free, Premium or Sales Navigator), not the HeyLead plan.
Privacy
AI calls — routed through HeyLead's backend or your own key
Cloud MCP — your data is processed server-side but never shared with third parties
Local mode — contacts and messages stay on your machine in a local SQLite database
Tool-call telemetry — when you are signed in, your HeyLead workspace records each tool call: the tool, whether it worked, the error type, how long it took and the app that called it (Claude, ChatGPT, Cursor…). Never the arguments, the results, a prospect's name, a link or a message. Turn it off with
heylead config telemetry offorDO_NOT_TRACK=1
Power users: Pass your own LLM key (Gemini/Claude/OpenAI) during setup to use your own AI. Completely optional.
Where secrets are kept
HeyLead keeps its backend token, the Unipile key and any LLM keys in the OS
credential store (macOS Keychain, Windows Credential Manager, Linux Secret
Service), or in an owner-only secrets.json in the HeyLead folder when none
exists. The config file holds ordinary settings only. To inspect an install
without seeing a secret:
heylead config get backend_urlprints one ordinary settingheylead api get /api/v1/campaignsmakes an authenticated read of the hosted API, credentials maskedheylead secrets statuslists which secrets are stored, never their valuesheylead secrets cleanmoves secrets out of an older config file, drops keys a cloud-sending install does not use, and deletes old config backup copies
Backend mode & env
When the MCP client talks to a HeyLead backend (e.g. heylead-api), the backend uses these environment variables. Operators running their own backend should set them as required.
Purpose | Example env vars |
LLM |
|
Search / crawl |
|
Auth / storage |
|
Optional | Feature flags, rate limits, logging — see backend repo |
For full backend configuration and deployment, see the heylead-api (or backend) repo and its docs.
Optional Dependencies
The base install covers all core features. For advanced ICP generation:
pip install heylead[icp] # Embeddings for RAG-powered ICP generation
pip install heylead[crawl] # Web crawling for company context ingestion
pip install heylead[all] # BothTroubleshooting
"heylead: command not found" or "uv: command not found"
Install uv first (curl -LsSf https://astral.sh/uv/install.sh | sh or brew install uv on Mac), then run uv tool install heylead and uv tool update-shell (or pass the full path from command -v heylead).
"MCP server not connecting" Restart your editor after adding the MCP server. In Cursor, check Settings > MCP — the server should show a green dot.
"Setup failed" or "LinkedIn not connected" Make sure you clicked "Connect" on the LinkedIn row of the sign-in page (dashboard: Settings → Connected accounts) and completed the LinkedIn login. Then run setup again.
Need help? Open an issue.
Publishing to PyPI (maintainers)
To make HeyLead available on PyPI (or to publish a new version):
Option A: Publish via GitHub Release (recommended)
One-time: Create a PyPI account and an API token. In your repo: Settings → Secrets and variables → Actions → add secret
PYPI_TOKENwith the token value.Bump version in
pyproject.toml(version = "0.2.4").Commit, push, then create a GitHub Release (tag e.g.
v0.2.4, release title optional). The workflow.github/workflows/publish.ymlruns on release and publishes to PyPI.
Option B: Publish manually
pip install build twine
python -m build # creates dist/
twine check dist/* # optional: validate
twine upload dist/* # prompts for PyPI username + password (use __token__ and your API token)After publishing, anyone can install it with uv tool install heylead and add it with claude mcp add heylead -- heylead (Cursor: command heylead).
For AI Agents
HeyLead is designed as an MCP-native tool — built for AI agents, not humans clicking buttons.
Install as MCP server (stdio):
{
"heylead": {
"command": "heylead",
"args": []
}
}OpenClaw: Add the same entry to your openclaw.json under mcp.servers.
Also available on ClawHub — search "HeyLead".
Sign in at heylead.dev and connect LinkedIn.
Capabilities: LinkedIn lead generation, cold outreach automation, ICP generation with buyer personas, voice-matched personalized messaging, multi-touch drip sequences, reply sentiment classification, engagement warm-ups, campaign analytics, and scheduled sending from the cloud with your laptop closed.
22 tools, from finding the right people to answering their replies: prospect discovery, ICP generation with buyer personas, connection invitations with a note, opening messages and follow-ups, warm-ups (profile views, follows, comments), reply handling and outcome tracking. Replies are read, classified and answered from the campaign's facts. A reply that could mean two things gets one short question back. A question your facts can't answer is held for you.
See AGENTS.md for the full agent integration guide.
Links
ClawHub (OpenClaw skill store)
License
MIT (code) — see LICENSE
Knowledge base and prompt configurations are proprietary.
Available Tools
22 toolsaccountADestructiveInspect
Switch the active LinkedIn account, unlink it, or connect an email account.
Which account is active decides who outreach is sent as. Listing them is
the accounts tool.
Args:
action: "switch", "switch_to", "unlink", "connect_email" or "refresh_tier".
account_id: The account id (switch_to).
Inputs: action (required), account_id. Returns: result
accountsARead-onlyInspect
List the LinkedIn accounts connected to this HeyLead. Changes nothing.
Args:
action: "list" (switching, unlinking and connecting email are the account tool).
Inputs: action. Returns: result
analyticsARead-onlyInspect
Campaign analytics — reports, comparisons, and exports.
Hosted accounts get a dashboard link and a snapshot card; relay the link,
because some clients show the image only to the model.
Args:
action: What to do:
"report" — Detailed analytics with outcomes, conversion rates, stale leads
"compare" — Compare 2+ campaigns side by side
"export" — Export campaign results as table, CSV, or JSON
campaign_id: Which campaign. Uses active if empty.
campaign_ids: Comma-separated IDs (for 'compare'). Compares all if empty.
format: Output format for 'export': 'table', 'csv', or 'json'.
Inputs: action, format, campaign_id, campaign_ids.
answer_inboxADestructiveInspect
Reply in a LinkedIn conversation, or approve or discard a drafted reply.
"reply" and "approve_draft" SEND a LinkedIn message and cannot be undone.
Reading the inbox is the inbox tool.
Args:
action: "reply", "approve_draft" or "discard_draft".
chat_id: Which conversation, or which draft.
text: What to send, or an edited version of the draft.
Inputs: action (required), text, chat_id. Returns: result
book_meetingADestructiveInspect
Book a meeting on your Google Calendar and invite a prospect.
Use this when a reply agrees to a call. Creates the event on the calendar
you connected, attaches a Google Meet link, and emails the attendee an
invitation.
Args:
attendee_email: Who to invite — the prospect who replied.
start: When it starts, ISO 8601, e.g. 2026-09-01T10:00:00.
duration_minutes: How long the meeting runs. Defaults to 30.
summary: Event title. Defaults to naming the attendee.
description: Optional agenda or notes included in the invitation.
Inputs: start (required), attendee_email (required), summary, description, +1 more. Returns: result
campaignADestructiveInspect
Launch, pause, resume, archive, delete or emergency-stop a campaign.
Every action here changes what happens next. Watching one run, or reading
its history, is campaign_status.
Args:
action: "launch", "pause", "resume", "archive", "delete",
"emergency_stop", "retry_failed", "repair_queue" or
"clear_coordinator_hold".
campaign_id: Which campaign. Uses the active one if empty.
confirm: Required by the actions that destroy work.
icp_hold_ok: launch and resume only: True after the user chose "launch
anyway" on a target the ICP research agent held.
Inputs: action (required), confirm, campaign_id, icp_hold_ok.
campaign_statusARead-onlyInspect
Watch a campaign run, show its plan, or read its status history. Changes nothing.
Args:
action: "monitor" (live progress; a draft is started with
campaign(action='launch')), "plan" (what happens after launch,
step by step: finding, warm-up, invitations, opening message,
follow-ups, replies, leads) or "status_history".
campaign_id: Which campaign. Uses the active one if empty.
Inputs: action, campaign_id.
check_repliesADestructiveInspect
Read new LinkedIn replies and record what they mean.
Fetches new messages, classifies each reply (positive, negative, question)
and lists the people worth answering first.
It sends nothing, but it does write: it marks an invitation accepted once
the person answers or connects, moves an outreach to replied, stores the
messages, and sets opted-out when someone asks not to be contacted, which
stops all future outreach to them. It is the only path that notices an
accepted invitation, so follow-ups depend on it having run.
Inputs: no named parameters. Returns: result
contactsARead-onlyInspect
Search, browse and read your contacts, and export them. Changes nothing.
Tagging, noting, re-staging and linking are update_contact.
Args:
action: "list", "search", "view", "stats", "export", "linkedin_search"
or "my_connections".
query: Search text for "search", "linkedin_search" and "my_connections".
For "linkedin_search" the query is passed to LinkedIn as KEYWORDS,
matched literally — a company name, a job title, a person's name, or a
combination such as 'Acme Corp CTO' or 'Jane Doe'. A natural-language
question ('who is the CTO of Acme?') is sent through unchanged and
usually comes back empty, so prefer keywords. Nothing is filtered out
locally. An empty result and a failed search are reported in different
words, so a "no matches" line means LinkedIn really returned nobody
rather than "the search broke". Results are capped at 25 per call
because every result costs a profile fetch and a LinkedIn read.
contact_id: Which contact (view).
lifecycle_stage: Filter by stage.
min_fit_score: Only those scoring at least this.
limit: How many.
format: "table", "csv" or "json" (export).
campaign_id: Filter by campaign.
connected_since: Only those connected on or after this date.
connected_before: Only those connected before this date.
Inputs: limit, query, action, format, +6 more. Returns: result
create_campaignADestructiveInspect
Create a LinkedIn outreach campaign from a natural language description.
Finds people on LinkedIn who match the description and saves a draft;
nothing is sent until campaign(action='launch'). Use it for sales
prospecting, recruiting and candidate sourcing, research and
user-interview recruitment, job-search networking, investor and partner
outreach, vendor scouting and event invitations.
Reaching people you can already name — an article author, a warm intro, a
speaker, one founder a customer mentioned — is the `people` argument: pass
their profile URLs and the campaign is seeded from exactly them, with no
LinkedIn search. Everything else is unchanged: draft until launch, rate
limits, sending window, opt-outs.
Describe who you need to reach and HeyLead finds them on LinkedIn:
customers (lead generation, B2B prospecting, cold outreach), candidates
(recruiting and sourcing), hiring managers and referrals (job search),
investors and partners, vendors, and user-interview or research participants.
On first campaign, project_brief is asked explicitly, in the goal's words
(see project_brief below) — a homepage alone is not enough.
Args:
target_description: Who to target (e.g., "CTOs at fintech startups",
"freelance UX designers in London", "yoga studio owners in California")
campaign_name: Optional name for the campaign.
icp_id: Optional ID of a saved ICP from generate_icp. If provided,
uses the saved ICP's enriched LinkedIn codes for precise targeting
instead of generating a new one.
company_context: Optional. Your website URL or 1-2 sentences about your
product/company. Copied into project_brief when project_brief is omitted.
project_brief: Optional. Full project paste the model sees, in the goal's
words. sell: what you offer and who it is for. job_search: the role,
the kind of company, what you bring. hire: the role, who fits it, what
it offers. partner: what you want from a partner, what you bring.
buy: what you need, by when and how much, what a vendor must confirm.
research: what you research, who you want to hear from, what you ask.
Required before launch, resume, or auto-send.
mode: Always autopilot (copilot mode was removed). Whether opening
DMs, follow-ups and answers to replies wait for a person is the WORKSPACE's approval mode,
not this: a hosted workspace that never chose sends them without
review; scheduler(action='approval_mode', mode='require_approval')
holds them (inspect(action='waiting') lists what waits).
company_url: Optional LinkedIn company URL for account-based targeting.
Searches for employees at that specific company matching the ICP.
Example: "https://www.linkedin.com/company/google"
voice_mode: "text_only". Messages are text.
connections_only: "on" to create a DM-only campaign targeting existing
LinkedIn connections. Skips invitations and warm-up — sends DMs
directly to people you're already connected with. Use when user says
"existing connections", "DM my network", "message my connections".
exclude_connections: ON BY DEFAULT for a new campaign: nobody who was
already a 1st-degree connection before this campaign started is
reached — refused at enrolment and skipped at send time rather than
DMed. People who accept this campaign's own invitation still get
the opener. Pass "off" to include existing connections (or turn it
off later in the campaign settings). Defaults off only for a
connections_only campaign. Use "off" when the user says
"include my existing connections"; the old "on" is still accepted
for "don't message my existing connections", "cold only",
"skip people I already know". Cannot be combined with
connections_only, which is its exact inverse.
people: LinkedIn profile URLs or public identifiers, comma or newline
separated (e.g. "linkedin.com/in/jane-doe, linkedin.com/in/john-doe").
The campaign is seeded from exactly these people: no LinkedIn
search runs, the goal <-> ICP audit is skipped (the audience is
stated, not inferred), a low ICP score does not drop anybody, and
discovery stays off so nothing tops the queue up with strangers.
Use it whenever the user names who to reach. Cannot be combined
with connections_only.
campaign_type: Prompt family: "outbound" (default) or "job_search".
job_search writes a job-search campaign: the invitation note and
the first DM may name the recipient's company and the role, use
one credible proof point at most and never list a CV. InMail is
not routed by this switch. Pair with connections_only="on" to
write to people the sender is already connected to.
force: True to create the campaign even when the goal <-> ICP audit
returns `mismatch` (the ICP holds no plausible buyer for the goal).
Leave False; a `partial` verdict never blocks, it only warns.
goal: What the campaign is for: "sell" (default), "job_search",
"hire", "partner", "buy" or "research". It sets campaign_type and
campaign_intent, picks whose profile the ICP describes, and picks
the fit question the goal <-> ICP audit asks. hire, partner and
research run on your project_brief until their message sets exist.
message_language: The language every message to a prospect is
written in: "en", "prospect" (their latest message, else their
LinkedIn profile's language, else English) or a code: uk, de, fr,
es, pl, pt, it. Never set this without asking the user. Empty
leaves it unset: messages go out in English, and launch asks once
when the brief is in another language.
Inputs: target_description (required), goal, mode, force, +11 more. Returns: result
edit_campaignADestructiveInspect
Change one running or drafted campaign's settings.
Pass only what you want to change; every field left empty keeps its current
value. Use it to rename a campaign, give it a booking link, feed it context
that makes the messages more specific (offerings, case studies, project
brief), say who it is for (campaign_intent, campaign_type), or set the
limits it sends under (volume, caps, follow-ups, business hours).
It edits settings only: it sends nothing and adds no prospects. A change
applies to the messages written from now on, not to ones already sent. Use
create_campaign for a new campaign, and campaign(action=...) to launch,
pause or stop one.
Args:
campaign_id: Which campaign to edit. Edits the first active campaign if empty.
name: New campaign name. Leave empty to keep current name.
mode: Only "autopilot" supported (copilot mode was removed). Whether
messages wait for a person is the workspace's approval mode:
scheduler(action="approval_mode").
booking_link_page_ok: Explicitly keep a website page as the booking link.
booking_link: Calendar/booking URL (e.g., "https://cal.com/you/15min").
Used in reply_to_prospect() for positive replies to suggest meetings.
offerings: What you offer (products, services, value props). Used in follow-up messages.
case_studies: Brief case studies or success stories. Used for social proof in messages.
social_proofs: Social proof (logos, metrics, testimonials). Used in follow-up messages.
campaign_preferences: Custom messaging preferences (tone, topics to avoid, etc.).
campaign_intent: Message stance: "sell", "buy", "partner", "recruit" or "research".
campaign_type: Prompt family: "outbound" (default) or "job_search".
job_search replaces the intent-specific invitation note and first
DM with ones that may name the company and the role, use one
credible proof point at most and never list a CV. InMail is not
routed by this switch, and campaign_intent still selects the
system prompt. Empty keeps the current value.
goal: What the campaign is for: "sell", "job_search", "hire",
"partner", "buy" or "research". Rewrites campaign_type and
campaign_intent to match. Empty keeps the current value.
project_brief: Full project paste the model sees. Required before launch,
resume, or auto-send.
offer_outcome: The Offer card's outcome: what changes for the reader,
in their words. Editing it unconfirms the card.
offer_how: The Offer card's how: what the sender does, said only in a
reply. Editing it unconfirms the card.
offer_proof: The Offer card's proof, used at most once per thread.
Editing it unconfirms the card.
offer_ask: The Offer card's one question. Editing it unconfirms the card.
offer_confirm: "on" confirms the Offer card; first messages resume.
product: Optional structured fact: product / what you buy or sell.
go_live: Optional structured fact: go-live date.
volume: Optional structured fact: volume model.
must_confirm: Optional comma-separated questions a vendor must confirm.
voice_mode: "text_only". Messages are text. Leave empty to keep current value.
enable_profile_views: View prospect profiles before following: "on" or "off".
enable_follows: Follow prospects before inviting: "on" or "off".
enable_endorsements: Endorse skills before inviting: "on" or "off".
enable_engagements: Comment/react on posts before inviting: "on" or "off".
enable_followups: Send follow-up DMs after connection: "on" or "off".
enable_auto_replies: "on" or "off". On: HeyLead answers replies itself,
or holds each answer in Approvals when the workspace reviews every
message. Off: HeyLead answers no reply and drafts none; every reply
waits for you in Needs attention, and no booking link or calendar
invite goes out. To read answers before they go, keep it on with
review mode, not off.
enable_invitations: Send connection invitations: "on" or "off".
When off, campaign only DMs existing connections (no invitations sent).
enable_discovery: Auto-find and enrol new prospects: "on" or "off".
exclude_connections: "on" to never message anyone who was already a
1st-degree connection before this campaign started — they are
refused at enrolment and skipped at send time instead of being
DMed. People who accept this campaign's own invitation still get
the opener. Turning it on turns connections_only off.
connections_only: "on" to target only your existing 1st-degree
connections (DM-only, no invitations). Turning it on turns
exclude_connections off.
Turn off for curated campaigns with a fixed, hand-picked list.
exclude_competitors: "on" to never first-touch people who work at
competing companies. Default on. Empty list excludes nobody
until research or competitor_companies names them.
competitor_companies: Comma-separated employer names to skip.
enable_reply_agent: Reply exception agent: "on" (act), "off", or
"observe". Empty keeps the current value. Unset defaults to act.
enable_strategist_replan_agent: Strategist replan: "on", "off", or
"observe". Empty keeps the current value.
enable_hot_lead_closer: Hot-lead closer: "on", "off", or "observe".
Empty keeps the current value. Unset defaults to observe. It
never sends a calendar invite from this computer: "on" notes a
call that is ready for you to book.
enable_coordinator_agent: Coordinator digest/hold: "on", "off", or
"observe". Empty keeps the current value.
engagement_mode: Engagement style: "auto" (30% react / 70% comment),
"comment_only", or "react_only".
max_followups: Max follow-up messages (1-5). 0 to keep current.
weekly_meeting_target: Meetings this campaign should book per week.
The daily report reads it as the Key Result and says whether the
campaign is on track. 0 means no goal this week; -1 keeps current.
followup_delay_days: Custom day intervals as comma-separated list
(e.g., "1,3,7,14"). Leave empty to keep current.
invite_note: Invitations carry a note: "on" (default) or "off". Off sends
the bare invitation; the first words are the DM a working day after
they accept.
withdraw_stale_invites: Withdraw an unanswered invitation on LinkedIn after
stale_invite_days: "on" (default) or "off".
stale_invite_days: Days before withdrawing stale invites (7-60). 0 to keep current.
inmail_fallback: Escalate quiet invitations with one InMail: "on" or "off".
Free tier sends only to Open Profile members (zero credits).
inmail_fallback_days: Quiet days before the InMail (1-60). 0 to keep current.
inmail_first_touch: InMail as first touch: "on" or "off". Unset follows inmail_fallback.
send_in_business_hours: Send only in business hours: "on" or "off". On by default
(weekdays 08:00-22:00 in your own timezone, or London when it is unknown,
unless the workspace set its own window).
active_days: Active send days as comma-separated numbers (0=Mon, 6=Sun).
E.g., "0,1,2,3,4" for weekdays. Leave empty to keep current.
message_language: The language every message to a prospect is
written in: "en", "prospect" (their latest message, else their
LinkedIn profile's language, else English) or a code: uk, de, fr,
es, pl, pt, it. Never set this without asking the user. Any
explicit value, "en" included, answers launch's language question.
Empty keeps the current value.
Inputs: goal, mode, name, volume, +50 more. Returns: result
generate_icpADestructiveInspect
Generate a rich Ideal Customer Profile with buyer personas.
The same profile describes whoever the user needs to reach: buyers,
candidates to recruit, research or user-interview participants, hiring
managers for a job search, investors or partners. Pass goal= so the
profile is of the right people.
Creates 2-4 ICP personas with pain points, fears, barriers,
LinkedIn search parameters, and confidence scores. The result
is saved and can be reused with create_campaign(icp_id=...).
Supports target audience analysis, customer segmentation, buyer persona
creation, ideal customer profiling, and B2B market research.
Args:
target_description: Who to target (e.g., "CTOs at fintech startups",
"freelance UX designers in London", "yoga studio owners in California")
company_context: Optional URL or text about your company/product.
Providing this makes the ICP more precise and evidence-backed.
focus_query: Optional focus (e.g., "enterprise segment only",
"focus on pain points around compliance")
decision_makers_only: Keep every persona's seniority to people who
hold budget authority — owner, cxo, vp, director. Default True.
Pass False only when the target really is individual contributors
(developers, designers, analysts); the ICP then keeps whatever
levels the description implies. Applies only to sell, partner
and buy; managers hire, so a job search keeps them.
goal: What the campaign is for, which decides whose profile this is:
"sell" (customers, the default), "job_search" (the people who hire for
or refer into the role), "hire" (candidates), "partner" (who can sign
a partnership or invest), "buy" (vendors), "research" (participants).
Always pass it; a job search with goal="sell" produces peers, not
hiring managers.
Inputs: target_description (required), goal, focus_query, company_context, +1 more. Returns: result
inboxARead-onlyInspect
Read any conversation in your LinkedIn inbox, and the drafted replies.
Replying, approving a draft and discarding one are answer_inbox.
Args:
action: "list", "read", "open_file" (save the files a person sent, to read them) or "comment_drafts".
chat_id: Which conversation.
name: Find the conversation by the person's name.
limit: How many to show.
Inputs: name, limit, action, chat_id. Returns: result
prospectADestructiveInspect
Skip a prospect, close one with an outcome, or dismiss it.
Someone who is not a fit is skipped, not closed: skip leaves them out of
this campaign and records nothing else. Close records a result and
needs an outcome; won and lost are for people HeyLead has written to.
Closing with outcome="opt_out" stops all future contact with that person.
Reading a prospect is prospect_view.
Args:
action: "skip", "close" or "dismiss".
outreach_id: Which outreach.
campaign_id: Which campaign. Uses the active one if empty.
outcome: "won", "lost" or "opt_out". Required for close; there is
no default.
reason: Why, recorded with the outcome.
meeting_link: The booked meeting, recorded with a won outcome.
confirm: Required where the action cannot be undone.
reason_code: Why, for skip and close: "not_a_fit", "negative_reply",
"asked_to_stop", "handled_elsewhere" or "other".
reason_note: A free note alongside the code.
deal_value: What the won deal is worth (close with outcome="won").
deal_currency: Three-letter code, e.g. USD. Defaults to the last deal's.
Inputs: action (required), reason, confirm, outcome, +7 more. Returns: result
prospect_viewARead-onlyInspect
Read the message thread with one prospect, or their timeline. Changes nothing.
Args:
action: "conversation" or "timeline".
outreach_id: Which outreach.
campaign_id: Which campaign. Uses the active one if empty.
Inputs: action, campaign_id, outreach_id. Returns: result
schedulerADestructiveInspect
Turn the autonomous scheduler on or off, or change how it runs.
Off means nothing sends until it is on again. Reading its state, logs,
activity or diagnostics is scheduler_status.
Args:
action: "toggle", "observe", "always_on", "backfill_cloud",
"send_from" or "report".
enabled: True to enable, False to disable (toggle, report).
cloud: Act on the hosted scheduler rather than this machine's.
host: "cloud" (send_from); a hosted account never sends from here.
hours: The reporting interval (report).
campaign_id: Which campaign (report).
Inputs: action (required), host, cloud, hours, +2 more. Returns: result
scheduler_statusARead-onlyInspect
Read the scheduler: status, logs, activity, diagnostics or the daily report.
Args:
action: "status", "logs", "activity" or "diagnostics". The daily
report is the scheduler tool: it switches reporting on and off.
cloud: Read the hosted scheduler rather than this machine's.
hours: Lookback window in hours.
event_type: Filter the log by event type.
campaign_id: Filter by campaign.
Inputs: cloud, hours, action, event_type, +1 more. Returns: result
send_messageADestructiveInspect
Send follow-ups, replies, or InMail to prospects.
Args:
action: What to do:
"followup" — Send a follow-up DM after connection accepted
"reply" — Reply to a prospect who has messaged you
"delete" — Delete a recently sent message (within 60 min on LinkedIn)
"inmail" — Send an InMail to a NON-connection. Preconditions
(each fails closed with no send): prospect has a
provider_id; they are NOT a 1st-degree connection
(use followup/DM for those); no InMail already sent
on this outreach; InMail credits remaining > 0.
A pending invitation to the same person is allowed
— that is the escalation path. Requires outreach_id.
campaign_id: Which campaign to send from. Uses active if empty.
outreach_id: Specific outreach to target. Required for inmail.
format: 'text'. For followup/reply.
text: For followup/reply on a hosted account: the message to send,
required; the cloud sends it as is in the outreach's LinkedIn
chat (the same route as answer_inbox reply). Without a backend
the message is generated. For delete: optionally pass a Unipile
message_id directly.
Inputs: text, action, format, campaign_id, +1 more. Returns: result
setup_profileADestructiveInspect
Set up HeyLead by connecting your LinkedIn account and analyzing your writing style.
REQUIRED for first-time users — must be called before any other tool.
This analyzes your LinkedIn profile, posts, and writing style to create
a "voice signature" so every outreach message sounds like YOU, not a bot.
Handles LinkedIn automation setup, SDR onboarding, account connection,
and voice analysis for personalized outreach.
First-time setup: sign in at https://heylead.dev/auth/login-url, click
'Connect' on the LinkedIn row, copy the setup message ('Copy' under
'Get Started'), then call this tool with the eyJ... token from it as
backend_jwt. No API keys needed on the hosted backend.
Args:
llm_api_key: Optional — only if you want to use your own AI key instead of the backend's.
llm_provider: Which AI to use if providing your own key: "gemini", "claude", or "openai".
backend_url: HeyLead Backend API URL. Leave empty — defaults to production server.
backend_jwt: Your authentication token from HeyLead.
Inputs: backend_jwt, backend_url, llm_api_key, llm_provider. Returns: result
show_statusARead-onlyInspect
Show your outreach dashboard — campaigns, stats, hot leads, account health.
The chat is the front door to your dashboard. View pipeline metrics, prospect
funnel, engagement rates, and campaign performance. Ask "how's my outreach?" anytime.
Hosted accounts get a dashboard link and a snapshot card; relay the link,
because some clients show the image only to the model.
Args:
campaign_id: Show stats for a specific campaign. Shows all if empty.
Inputs: campaign_id.
suggest_next_actionARead-onlyInspect
Suggest the best next action for your outreach.
Hosted accounts get a dashboard link and a snapshot card; relay the link,
because some clients show the image only to the model.
Analyzes all active campaigns and recommends what to do next,
prioritized by impact: hot leads first, then pending approvals,
follow-ups, engagement warm-ups, and new invitations.
Args:
campaign_id: Focus on a specific campaign. Analyzes all active if empty.
Inputs: campaign_id.
update_contactADestructiveInspect
Change one contact: tag it, note it, move its stage, or link it to a campaign.
Reading and searching contacts is the contacts tool.
Args:
action: "tag", "note", "stage", "link" or "enrich".
contact_id: Which contact.
lifecycle_stage: The stage to move it to.
tag: The tag to add.
note: The note to record.
campaign_id: Which campaign (link).
match: How to match when linking.
dry_run: Show what would change without changing it.
Inputs: action (required), tag, note, match, +4 more. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 22 tools
Several tools answer the same question: show_status, analytics, campaign_status and suggest_next_action all report on outreach health, and reply flows are split across inbox, check_replies, answer_inbox and send_message(action='reply'). The descriptions cross-reference each other explicitly ("Reading the inbox is the inbox tool"), which mitigates most misselection, but the boundaries are not self-evident from names alone.
Everything is snake_case, but the set mixes imperative verb_noun names (create_campaign, send_message, update_contact, book_meeting) with bare-noun tools that take an action parameter (campaign, prospect, scheduler, inbox, contacts). Near-identical pairs like account/accounts and campaign/campaign_status make the pattern harder to predict.
22 tools is heavy for what is a fairly coherent LinkedIn outreach domain; each tool maps to a real operation, but the surface is split into many narrow read tools (inbox, contacts, accounts, campaign_status, scheduler_status, analytics, show_status) that could be consolidated. It sits at the top of the reasonable range rather than being well-scoped.
The surface covers the full lifecycle: setup, ICP generation, campaign create/edit/launch/pause/delete, prospect management, contacts, inbox reading and replying, messaging, meeting booking, scheduling and analytics. Minor gaps remain (no dedicated list_campaigns, no explicit contact create/delete), but agents can work around them via existing tools.
Maintenance
Related MCP Connectors
LinkedIn outreach in Claude or ChatGPT: campaign stats, replies, leads, drafts. Never sends.
Run LinkedIn outreach from your AI chat: find leads, launch campaigns, send, and reply.
Find LinkedIn prospects, draft outreach, and run human-paced campaigns from your AI agent.
Run governed LinkedIn outreach from your AI assistant: campaigns, replies, stats and 30+ languages.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables sending LinkedIn messages and invitations, reading conversations, and managing outreach through natural language by automating a real LinkedIn session via a Chrome extension.18 npmMIT
- AlicenseNot gradedqualityCmaintenanceConnects AI assistants to LinkedIn outreach, enabling lead finding, campaign management, messaging, and analytics through natural language.MIT
- AlicenseAqualityBmaintenanceLinkedIn prospect intelligence inside Claude. It automates the research behind outreach, not the sending: scrapes posts, profiles and comment threads into a local SQLite database, classifies what a person or company actually posts about, mines comment threads for people already describing your problem, and surfaces uncommon commonalities worth opening with.1834MIT
- AlicenseNot gradedqualityBmaintenanceLinkedIn from the user's own logged-in Chrome session for AI agents: search, profiles, post engagers, company employees, lists, sequences, inbox and a Research Pack, with hard caps and a human approval queue. Local-first, no headless browser.15MIT