li — The LinkedIn CLI for agents

Stop opening LinkedIn. Let your agents do it.

npm install -g @peggyrayzis/li
li whoami

Built by @peggyrayzis of scale.dev — marketing & GTM for devtools and AI founders. Working on something cool? Reach out at [email protected].

Why

I was tired of building growth workflows that depended on Clay — which doesn't have an API. Then I saw Bird (a CLI for Twitter) and thought: LinkedIn needs this.

So I built li. Cookie auth, LinkedIn's internal Voyager API, structured JSON output. Now my agents handle prospecting, connection monitoring, and message triage without me ever opening a browser tab.

Quick Start

Option A: Auto-extract cookies from Chrome or Safari

li check

On macOS, li can read your LinkedIn cookies directly from Chrome or Safari. If it works, you're done. Fair warning — cookie extraction can be flaky depending on your OS and browser version. If it doesn't work, Option B is reliable.

Option B: Set cookies manually

# Open Chrome DevTools → Application → Cookies → linkedin.com
# Copy li_at and JSESSIONID values

export LINKEDIN_LI_AT="your-li-at-value"
export LINKEDIN_JSESSIONID="your-jsessionid-value"
li check

Verify

li whoami          # See your profile
li connections     # List your connections
li messages        # Check recent conversations

Commands

v0.1 is read-only. Write commands (connect, send, invites accept) ship in v0.2.

Identity

li whoami                         # Your profile + network counts
li check                          # Validate session, show credential source

Profiles

li profile peggyrayzis            # View by username
li profile linkedin.com/in/user   # View by URL

Connections

li connections                    # List connections (default: 20)
li connections -n 50              # Show 50
li connections --all              # Fetch all (paginated)
li connections --of peggyrayzis   # View someone else's connections

Note: li connections --of ... depends on LinkedIn search pagination behavior. LinkedIn may cap results (often around ~20) even when more pages are requested.

Search

li search --query "founder"               # Search profiles
li search --query "founder" -n 10         # Return up to 10
li search --query "founder" --all         # Request all pages (still capped at 50)
li search --query "founder" --fast        # Faster fetch mode
li search --query "founder" --json        # Machine-readable output

Supported flags: --json, --fast, -n/--count, --all.

Temporary MVP limit: search results are hard-capped at 50 profiles. --all still stops at 50.

MVP JSON output for li search --query ... --json uses these top-level keys:

{
  "query": "founder",
  "limitApplied": 50,
  "connections": [],
  "paging": {}
}

Invitations

li invites                        # List pending invitations

Messages

li messages                       # List recent conversations
li messages read CONV_ID          # Read a thread
li messages read CONV_ID -n 50   # Last 50 messages

Every command supports --json for piping to jq or feeding to agents.

Agent Workflows

li is built for agents. Pipe --json output into whatever you want.

Find CTOs in your network

li connections --all --json | \
  jq -r '.connections[] | select(.headline | test("CTO"; "i")) | .username'

Check unread messages

li messages --json | jq '.conversations[] | select(.unreadCount > 0)'

Find CTOs in a connection's network

li connections --of peggyrayzis --all --json | \
  jq -r '.connections[] | select(.headline | test("CTO"; "i")) | .username'

Export connections to CSV

li connections --all --json | \
  jq -r '.connections[] | [.firstName, .lastName, .headline, .username] | @csv'

Global Options

| Option | Description | |--------|-------------| | --json | Structured JSON output | | --li-at <token> | Override li_at cookie | | --jsessionid <token> | Override JSESSIONID | | --cookie-source <src> | chrome, safari, none, or auto (default) | | --no-progress | Suppress progress output | | -h, --help | Help | | -V, --version | Version |

Environment Variables

| Variable | Description | |----------|-------------| | LINKEDIN_LI_AT | Your li_at session cookie | | LINKEDIN_JSESSIONID | Your JSESSIONID cookie (CSRF token) |

Requirements

• Node.js >= 22
• A valid LinkedIn session (logged in via Chrome)

Rate Limiting

LinkedIn is aggressive about bot detection. li enforces minimum 500ms between requests with exponential backoff on 429s. For bulk operations, add your own delays between commands.

Programmatic API

A library API for using li in Node.js applications is planned for v0.2.

Contributing

PRs welcome. Here's the setup:

git clone https://github.com/peggyrayzis/li.git
cd li
pnpm install
pnpm test        # Run tests
pnpm lint        # Check with Biome
pnpm run build   # Build with tsup

This project uses TDD — write tests first, then implement. Tests live in tests/unit/ and fixtures in tests/fixtures/. Run pnpm test to confirm everything passes before opening a PR.

A few ground rules:

• No secrets in code. Real cookie values, API keys, or PII will be rejected.
• No external LinkedIn API libraries. We own the Voyager calls directly.
• v0.1 is read-only. Write commands (send, connect, invites accept) are deferred to v0.2.
• Open an issue first for large changes so we can discuss the approach.

Releasing

• Add a changeset for user-facing updates: pnpm changeset
• Merge feature PRs to main.
• Cut a release PR manually: bash scripts/cut-release-pr.sh
• Merge the generated chore: release PR.
• Run the Release workflow manually (workflow_dispatch) to publish npm + tag + GitHub release notes.

Disclaimer

This tool uses LinkedIn's internal Voyager API with cookie authentication for personal use and agent-powered workflows on your own account. LinkedIn can change their API at any time. Aggressive automation may result in account restrictions. Not affiliated with or endorsed by LinkedIn.

License

MIT