DenchClaw Zero to CRM: Full Setup in Under an Hour
Get DenchClaw running as your full CRM in under an hour. Step-by-step setup guide covering install, objects, fields, and your first pipeline.
DenchClaw Zero to CRM: Full Setup in Under an Hour
You can go from zero to a fully working CRM with DenchClaw in under an hour. I've done it dozens of times, and this guide is the fastest path I know.
By the end of this, you'll have DenchClaw installed, a contacts database running, a sales pipeline configured, and your first AI-assisted workflow active. No cloud accounts, no credit card, no support tickets.
Step 1: Install DenchClaw (5 minutes)#
Open your terminal and run:
npx denchclawThat's it. The installer:
- Bootstraps an OpenClaw profile called
denchon port 19001 - Initializes a local DuckDB database
- Launches the web frontend at
localhost:3100 - Copies your Chrome profile so you're already logged in everywhere
Add localhost:3100 to your Dock as a PWA (in Chrome: three-dot menu → More Tools → Create Shortcut → Open as window). It'll feel like a native app.
While it's installing, connect a messaging channel. Telegram is the fastest: create a bot via @BotFather, get the token, and paste it into DenchClaw settings. Now you can talk to your CRM from your phone.
Step 2: Create Your First Object — People (10 minutes)#
DenchClaw stores everything in Objects — think tables in a CRM. Let's start with people.
In the web chat or Telegram, type:
Create a People object with fields: Full Name (text), Email (email), Phone (phone), Company (text), Status (enum: Lead/Prospect/Customer/Churned), Last Contacted (date), Notes (richtext)
The AI will create the object, set up the fields, and generate a v_people pivot view in DuckDB. You'll see it appear in your sidebar immediately.
If you prefer manual setup, navigate to Objects → New Object in the sidebar.
Step 3: Create a Companies Object (5 minutes)#
Create a Companies object with fields: Company Name (text), Website (url), Industry (enum: SaaS/Fintech/Healthcare/Ecommerce/Other), Size (enum: 1-10/11-50/51-200/200+), ARR (number), Primary Contact (relation to People)
Now link people to companies. Go back to your People object and add:
Add a Company relation field to People, linking to Companies
The AI adds a many_to_one relation. Now each person has a linked company, and you can query across both.
Step 4: Set Up a Sales Pipeline (10 minutes)#
Create a Deals object:
Create a Deals object with fields: Deal Name (text), Value (number), Stage (enum: Discovery/Proposal/Negotiation/Closed Won/Closed Lost), Close Date (date), Company (relation to Companies), Owner (text), Notes (richtext)
Switch to Kanban view in the Deals object — each stage becomes a column. Drag deals between stages, or tell the AI:
Move "Acme Corp deal" to Negotiation stage
For your pipeline dashboard, ask:
Show me total pipeline value by stage as a bar chart
A Chart.js visualization renders inline. Save it as a report and it appears in your sidebar.
Step 5: Import Your Existing Contacts (10 minutes)#
From CSV: Navigate to Import or tell the AI:
Import this CSV into my People object [attach file]
The AI maps columns to fields and handles mismatches. It'll ask you to confirm before inserting.
From HubSpot: If you have a HubSpot account, say:
Import my HubSpot contacts
DenchClaw opens your browser (you're already logged in), exports the data, and maps it to your objects. See the full HubSpot import guide for details.
From Google Sheets:
Import contacts from my "Sales Q1" Google Sheet
Step 6: Add Your First AI Workflow (10 minutes)#
This is where DenchClaw separates from any other CRM.
Create an Action field on your People object:
Add an action field called "Enrich" to People that fetches company info from the web and updates the Company, Industry, and Size fields
Now every person in your database has an Enrich button. Click it, and the AI looks up their company on the web and fills in the fields. No API key required.
Set up a follow-up reminder:
When I mark a deal as Closed Won, remind me to schedule an onboarding call within 3 days
Create a saved filter view:
Create a view called "Hot Leads" showing people with Status=Lead, Last Contacted more than 14 days ago, sorted by Company
This view now appears in your sidebar as a permanent filter. Every morning, open Hot Leads and work down the list.
Step 7: Connect Your Calendar and Email (10 minutes)#
Tell the AI:
Connect my Google Calendar and Gmail
This uses the gog skill. Once connected:
- DenchClaw can add follow-up reminders to your calendar
- You can ask "what meetings do I have today?" from Telegram
- Email drafts can be generated from CRM context: "Draft a follow-up to Sarah Chen referencing our call last Tuesday"
What You Now Have#
After this hour, you have:
- ✅ A People database with full contact history
- ✅ A Companies database linked to people
- ✅ A Deals pipeline in Kanban view
- ✅ Imported contacts from your existing tools
- ✅ An AI enrichment workflow
- ✅ Saved filter views for daily use
- ✅ Calendar and email integration
- ✅ Mobile access via Telegram
Everything lives on your machine, in plain DuckDB files you can back up, version, or query directly. No vendor lock-in.
Next Steps#
Once you're comfortable with the basics:
- Build a custom analytics app — Chart.js dashboards on your CRM data
- Set up automated outreach — Email sequences triggered from your pipeline
- Power user guide — Advanced tips for getting more out of DenchClaw
For the full feature overview, see what DenchClaw is and how it works.
Frequently Asked Questions#
How long does the full DenchClaw setup actually take?#
For a basic People + Deals setup, under 30 minutes. Adding email, calendar, and import workflows takes the full hour. Complex team configurations can take a few hours.
Do I need to know SQL or code to use DenchClaw?#
No. The AI handles all DuckDB queries behind the scenes. You talk to it in plain English. If you want to write your own queries, you can — but it's entirely optional.
Can I use DenchClaw without a Telegram bot?#
Yes. The web chat at localhost:3100 works standalone. Telegram is the most convenient mobile interface, but it's optional.
What happens to my data if I stop using DenchClaw?#
Your data lives in a .duckdb file on your machine. You keep it. Open it with any DuckDB client, export it to CSV, or query it with Python. Zero lock-in.
Can I run DenchClaw on Linux or Windows?#
Yes. npx denchclaw works on macOS, Linux, and Windows. The browser profile copy feature is most seamless on macOS, but the rest works cross-platform.
Ready to try DenchClaw? Install in one command: npx denchclaw. Full setup guide →
