DenchClaw for Networking: Never Forget a Connection Again

DenchClaw is a local-first AI CRM that turns your professional network from a scattered list of LinkedIn connections into a structured, queryable relationship system. With it, you'll never forget to follow up with a key contact, never lose the context from your last conversation, and always know exactly who you should be talking to. Here's how to set it up for serious networking.

What is DenchClaw? — local-first, open-source CRM backed by YC S24. Runs on your machine in DuckDB. Free.

The networking problem that no app solves well#

Most people's professional network management strategy is: connect on LinkedIn, hope LinkedIn's algorithm reminds you they exist, occasionally notice they changed jobs, occasionally reach out when you need something.

This is passive networking. And it produces passive results.

Active networking means:

• Knowing who matters to you and why
• Reaching out before you need anything
• Remembering personal context (what they're working on, what they care about)
• Following up consistently
• Building reciprocal relationships, not just extracting value when convenient

The tools that exist for this are either too lightweight (contacts apps, LinkedIn) or too heavy (enterprise CRMs built for sales teams). DenchClaw hits the right level: structured enough to be queryable and actionable, flexible enough to capture the nuance of real relationships.

See also building a second brain with DenchClaw for how networking fits into a broader personal knowledge management system.

Step 1: Install DenchClaw#

npx denchclaw

Frontend at localhost:3100. Everything runs locally — your network data stays on your machine.

Step 2: Set up your Contacts object#

Create a Contacts object with these fields:

The last_contact_date and follow_up_date fields are the heartbeat of your networking system.

Step 3: Import your existing contacts#

From LinkedIn:

1. Go to LinkedIn → Settings → Data Privacy → Get a copy of your data
2. Download Connections.csv
3. Import into DenchClaw: open Contacts → Import → CSV
4. Map: First Name + Last Name → name, Email Address → email, Company → company, Position → role

LinkedIn's export includes connection date but not any interaction history. That's fine — you'll build that going forward.

From Google Contacts:

1. Go to contacts.google.com → Export → Google CSV
2. Import into DenchClaw, map fields accordingly

Manual entry: For your top 50 relationships, add them manually. Take 10 minutes to fill in the context fields. These are the relationships worth the extra effort.

Step 4: Set up your "who needs attention" view#

This is the most important view in your networking system.

In DenchClaw Table view, set up a filter and sort:

• Sort by: last_contact_date ascending (oldest contact at top)
• Filter: relationship_strength = Close or Warm

This shows you, in one view, exactly who you should be reaching out to. No manual thinking required.

For a more precise approach, set up a DuckDB-backed view:

-- Contacts by priority: overdue reaches based on relationship strength
SELECT name, company, role, last_contact_date, relationship_strength,
  DATEDIFF('day', last_contact_date, CURRENT_DATE) as days_since_contact,
  CASE 
    WHEN relationship_strength = 'Close' AND DATEDIFF('day', last_contact_date, CURRENT_DATE) > 30 THEN 'OVERDUE'
    WHEN relationship_strength = 'Warm' AND DATEDIFF('day', last_contact_date, CURRENT_DATE) > 60 THEN 'OVERDUE'
    WHEN relationship_strength = 'Cold' AND DATEDIFF('day', last_contact_date, CURRENT_DATE) > 90 THEN 'OVERDUE'
    ELSE 'OK'
  END as status
FROM v_contacts
WHERE relationship_strength != 'Dormant'
ORDER BY days_since_contact DESC;

This surfaces everyone who's overdue for contact, calibrated by how close the relationship is.

Step 5: Use entry documents for relationship context#

Every contact in DenchClaw has an entry document — a markdown file for free-form notes. This is where you build the context that makes networking feel genuine rather than transactional.

For each important contact, maintain a living document:

# Sarah Chen
 
## About
CTO at BuilderAI. Met at YC S24 batch kickoff. She was working on 
AI-native developer tooling. Super smart on distributed systems.
 
## What she cares about
- Hiring exceptional engineers (mentioned it as her #1 challenge)
- ML infrastructure efficiency
- Building in the open — strong OSS contributor
 
## Our history
- Mar 2025: Met at YC kickoff, talked for 20 min about DuckDB vs Postgres
- May 2025: Coffee chat in SF. She introduced me to two potential hires.
- Sep 2025: Sent her a link to the DuckDB 1.0 release, she loved it
- Jan 2026: Brief catch-up at a conference
 
## What I can offer her
- Engineering hire referrals (she's always hiring)
- Intros to ML infra founders
- Occasional DuckDB/database advice
 
## Open threads
- [ ] She mentioned wanting to connect with ML infra investors — reach out
- [ ] Follow up on BuilderAI Series A progress

When you sit down to reach out to Sarah, you open her document and instantly have full context. You're not starting from zero — you're continuing a conversation.

Step 6: Log interactions#

Create an Interactions object to log every meaningful touchpoint:

Fields: date, type (call, email, meeting, conference, dm), contact (relation), summary, follow-up, follow-up_date.

After every substantive interaction, log it. This takes 2 minutes. Over a year, it builds an invaluable record.

-- Your most active networking contacts this year
SELECT c.name, c.company, COUNT(i.id) as interactions
FROM v_contacts c
JOIN v_interactions i ON i.contact_id = c.id
WHERE i.date >= '2025-01-01'
GROUP BY c.name, c.company
ORDER BY interactions DESC
LIMIT 20;

Step 7: Use Calendar view for follow-up dates#

Switch to Calendar view on your Contacts object, using follow_up_date as the date field.

Now you see every follow-up as a calendar event. Set follow-up dates liberally — whenever you have a conversation, decide when you'll next check in and set the date. Check your Calendar view weekly and work through the list.

This is the behavior that separates people who maintain strong networks from people who let relationships go dormant.

Step 8: Leverage the browser agent for research#

Before reaching out to someone you haven't talked to in a while, context is everything. DenchClaw's browser agent copies your Chrome profile — already logged in everywhere. Use it to:

• Check their LinkedIn for recent job changes, posts, or milestones
• Look up recent news about their company
• Find any recent talks, articles, or content they've published

An automated "pre-contact brief" action field can pull this context and append it to their entry document before you reach out.

Networking categories and how to segment them#

Not all networking is the same. Set up categories in your category select field:

• Investors — people who could fund you or your work
• Co-founders/partners — people you might build with
• Customers — people who use or might use what you're building
• Advisors — people you go to for specific expertise
• Operators — people doing the same kind of work as you
• Community — people you're connected to through shared identity (alumni, events, etc.)
• Personal — friends, family, personal connections

Segmenting this way lets you think clearly about what you're investing in. A healthy network needs diversity across categories. If your whole network is investors, you're missing the operators who can give you tactical feedback. If it's all operators, you may be lacking the investors and advisors who open different doors.

The best personal CRM in 2026 comparison#

DenchClaw competes in the personal CRM space against tools like Clay, Folk, Monica, and Dex. The differentiators for networking specifically:

• Local storage — your network is your most valuable professional asset; it shouldn't be in someone else's database
• Full SQL access — query your network in ways no UI exposes
• Entry documents — rich per-contact context that goes beyond any field set
• AI action fields — automate research, draft outreach, generate briefings
• Free — no subscription that makes you deprioritize adding contacts

FAQ#

How many contacts should I actively track? Most people have 50-200 relationships that genuinely matter to their professional life. Don't try to actively maintain 1,000 contacts — that's a broadcast list, not a network. Go deep with fewer people.

What's the best way to reconnect with someone I haven't talked to in years? Keep it simple and honest. "Hey, I was thinking about [specific thing related to them] and wanted to reconnect. What are you working on these days?" No need to apologize for the gap. People appreciate genuine outreach regardless of timing.

How do I avoid seeming transactional when I reach out? The entry document helps here. If you have context on what they care about — their current projects, their challenges — lead with something genuinely useful to them. Give before you ask.

Can DenchClaw remind me to reach out to people? Yes — use the follow_up_date field and Calendar view. You can also build a daily briefing action that queries overdue follow-ups and presents them when you open DenchClaw.

Should I add people I've only met once to DenchClaw? If the meeting was substantive and you can imagine a reason to reach out again, yes. A brief entry with a couple notes about how you met and what they do costs 5 minutes and might pay off years later.

Ready to try DenchClaw? Install in one command: npx denchclaw. Full setup guide →