Skip to content
@crmkit

CRM Kit

AI-first CRM for ChatGPT, Claude and others

crmkit

crmkit

Your CRM is a file your agent reads.

Website · Documentation · Skills


crmkit is an agent-first CRM, built for AI agents - ChatGPT, Claude, Cursor - to drive directly. The agent is the interface: it loads a one-page manual and operates the CRM over plain HTTP. Headless by design (no UI). Responses are grepable plain text by default; JSON on request.

Fastest way in - paste this into Claude, Claude Code, or any agent that can fetch a URL, and it sets itself up:

Read https://api.crmkit.ai/start and follow it to get me set up with crmkit.

Prefer raw HTTP? Drive it directly:

# 1. begin login - a 6-digit code is emailed
curl -s -X POST https://api.crmkit.ai/auth/request -d '{"email":"you@example.com"}'

# 2. verify the code to get a token
curl -s -X POST https://api.crmkit.ai/auth/verify -d '{"email":"you@example.com","code":"123456"}'

# 3. operate the CRM
curl -s -X POST https://api.crmkit.ai/contacts \
  -H 'Authorization: Bearer ck_…' \
  -d '{"name":"Jane Doe","email":"jane@acme.com","stage":"lead"}'

Plain-text, grepable output - one labeled line per record:

contact_k7m2q  name="Jane Doe"  email=jane@acme.com  stage=lead  updated=2026-06-04T09:13Z
# 1 contact(s)

Why crmkit exists

A CRM looks like something an agent could just keep in a few markdown files or a scratch database table. That holds right up until it doesn't. Underneath, a CRM is a pile of small problems that are each easy to get subtly wrong and painful to fix once real data has piled up:

  • stable IDs that survive renames and re-imports,
  • relations between contacts, companies, and deals that stay consistent,
  • search, filtering, and pagination that hold up as records grow,
  • dedupe, validation, and safe concurrent writes,
  • history, audit, and per-tenant limits,
  • schema changes you can apply to a live database without losing data.

And a CRM is rarely driven by one agent. Several agents - and the teammates alongside them - work the same records at the same time, so you need one shared system of record: consistent reads, safe concurrent writes, and a clear trail of which agent or member did what. Files and ad-hoc tables fall apart exactly here.

crmkit is that core - one well-maintained, battle-tested foundation that many agents and deployments share, so you build on solved problems instead of rediscovering them ad hoc in every project.

Use cases

crmkit is useful anywhere an agent needs to remember people, organizations, opportunities, and follow-up work in one shared system:

  • Sales CRM - track leads, accounts, deals, activities, next steps, and pipeline movement without building a dashboard first.
  • Personal contact management - keep notes, reminders, relationship context, and follow-ups for your own network.
  • Customer support - record customers, companies, conversations, issues, escalations, and renewal or upsell opportunities.
  • Market monitoring - track funded companies, competitors, target accounts, hiring signals, product launches, and other entities an agent watches over time.
  • Fundraising - manage investors, intros, conversations, diligence status, commitments, and follow-up tasks across a raise.
  • Events & conferences - track the conferences, trade shows, and webinars you work, the contacts and companies you meet at each, and the follow-up tasks and deals that come out of them.

Example: ChatBotKit's AI Market Bot

  • a market-research and competitive-intelligence agent - is the market-monitoring use case above, made real.

What you get

  • No UI to build or maintain - the agent is the front end.
  • Plain text first - token-cheap, grepable, JSON on Accept.
  • Instructive errors - every failure tells the agent what to do next.
  • OTP auth, bearer tokens - email a code, paste it, get a token.
  • Single static Go binary - CGO_ENABLED=0, SQLite embedded, deploy as one file.
  • Extensible - every record carries a free-form custom object.

Skills

Drop-in agent skills - digest, import, backup, inbox-sync - for coding agents (on ChatGPT/Claude.ai, add the MCP connector at https://api.crmkit.ai/mcp instead). Install the set with one command (Claude Code, Cursor, and 40+ agents via npx skills):

npx skills add crmkit/skills

Or download the release zip and unzip into your skills directory:

curl -fsSL https://github.com/crmkit/skills/releases/latest/download/crmkit-skills.zip -o /tmp/crmkit-skills.zip \
  && unzip -o /tmp/crmkit-skills.zip -d ~/.claude/skills/

Each recipe needs CRMKIT_BASE_URL and a crmkit token (POST /auth/request/auth/verify).

Repositories

Repo Description
crmkit Server (crmkitd), CLI, and documentation
skills Agent skill definitions

Related

  • zot - an autonomous coding agent you don't babysit.

Pinned Loading

  1. crmkit crmkit Public

    AI-first CRM for ChatGPT, Claude and others

    Go 26 2

  2. skills skills Public

    Shell 3

Repositories

Showing 3 of 3 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…