Community Spotlight: How Teams Are Using DenchClaw

Since the Show HN launch hit 147 points and Garry Tan shared DenchClaw on X, we've heard from hundreds of people using it in ways the team didn't fully anticipate. Some are running it as a personal CRM for investor relations. Others have built entire client management systems for their agencies. A few are using it as a local-first alternative to Notion that actually talks back.

This is the first in an ongoing series highlighting how the DenchClaw community is putting it to work. These are real use cases from real users — and they're a better introduction to what DenchClaw can do than any feature list.

The Investor-Facing Founder#

Who: A solo founder in YC S25, pre-revenue, fundraising.

The problem: Managing 60+ investor relationships across warm intros, cold outreach, meetings, and follow-ups — without the investor seeing a polished CRM view that feels transactional.

The setup: A custom investors object with fields for fund name, partner name, stage focus, check size, intro source, and relationship warmth. A kanban pipeline with stages: Warm Intro → First Meeting → Partner Meeting → Term Sheet → Passed.

The workflow: After every call, the founder dictates notes to DenchClaw via Telegram while walking back to the office. The agent parses the note, updates the relevant investor entry, and sets a follow-up reminder. Before the next call, the founder asks: "Summarize my last 3 touchpoints with Sequoia's Jenny Shen and what we discussed" — the agent pulls the full history.

What they said: "I've tried Affinity and Folk. They're great tools but they're enterprise CRMs pretending to be personal tools. DenchClaw is actually mine — my data, my schema, my notes. It doesn't feel like I'm feeding someone else's product."

The Two-Person Agency#

Who: A design agency in Berlin, 2 founders, ~15 active client projects.

The problem: Client emails, Figma links, feedback cycles, invoices, and relationship notes all scattered across inboxes, Slack channels, and sticky notes.

The setup: Three objects — clients (companies), contacts (people at client companies), and projects (individual engagements). Each project links to a client and has a status pipeline: Scoping → Active → Review → Done → Invoiced.

The projects object includes: project type, budget, start date, deadline, deliverables (text list), Figma link, invoice status.

The workflow: Every Monday morning, they ask DenchClaw: "What projects are due this week, and which invoices are overdue?" The agent queries both objects and returns a prioritized list. Client-facing deliverables get drafted with the agent's help, reviewed, and sent directly via Gmail through the gog skill.

What they said: "We spent 3 hours setting up the objects. That was it. Now we actually know where every project stands on a Monday morning, which has never been true before."

The Open Source Maintainer#

Who: An OSS developer maintaining three popular GitHub libraries, 4,000+ GitHub stars combined.

The problem: Managing sponsor relationships, bug reporter follow-ups, contributor thanks, and conference speaking requests — all while staying focused on the code.

The setup: A sponsors object (companies and individuals funding the work), a contributors object, and a requests object for speaking, consulting, and collaboration asks. Action fields on sponsors send thank-you emails via Gmail when clicked.

The workflow: New GitHub issues mentioning "sponsorship" or "collaboration" get piped to DenchClaw via a GitHub webhook and land in the requests object automatically. The agent handles initial responses while the maintainer reviews the queue once a day.

What they said: "The part that surprised me was the GitHub integration. I didn't expect it to work as smoothly as it did — new collaboration requests show up in my CRM before I've even opened GitHub."

The Bootstrapped SaaS Team (3 People)#

Who: A 3-person SaaS startup, bootstrapped, ~$15k MRR.

The problem: Three founders, all wearing CS/sales/product hats simultaneously. Deals falling through because nobody updated Pipedrive after a call.

The setup: Migrated from Pipedrive to DenchClaw in a weekend. Same core objects — people, companies, deals — but now the AI agent maintains the records. Each founder talks to DenchClaw on their phone via Telegram. After any customer interaction, they send a voice note or quick text with what happened.

The workflow: The agent listens, identifies which company and deal is being referenced (it knows their entire CRM history), and updates the relevant record. No manual data entry. Weekly, the agent sends a deal status summary to the team Telegram channel.

What they said: "We went from 60% data quality in Pipedrive (everyone entered what they felt like entering) to basically 100% in DenchClaw. Because talking to it is actually easier than typing into Pipedrive, we all actually do it."

The Independent Consultant#

Who: An independent strategy consultant, solo, ~12 active client relationships.

The problem: Consulting relationships are long — years, sometimes decades. Standard CRMs feel too transactional. Keeping track of what matters (their kids' names, the initiative they care about, when their fiscal year ends) requires a tool that treats context as first-class.

The setup: A clients object with rich entry documents per client. Each document is a running log of everything — meeting notes, personal details shared in conversation, strategic context, what they're worried about, what they're proud of. The CRM fields track the contract and commercial facts; the document captures the relationship.

The workflow: After any interaction — a call, a LinkedIn message, a coffee — the consultant adds a note to the client's document via Telegram. The agent timestamps it and appends it to the right section. Before the next engagement, they ask: "What's the context on [Client]? Last 3 interactions and anything important I should remember."

What they said: "I've tried keeping this in Notion, in Apple Notes, in Bear. They're all the same problem: searching is a nightmare and nothing is connected. In DenchClaw, my notes are connected to the person and the context is always one question away."

The Startup Ecosystem Builder#

Who: A community manager at a regional startup accelerator, managing 80+ portfolio companies.

The problem: Tracking 80 portfolio companies across different stages, with different needs, different check-in cadences, and different support requests — without a dedicated CRM team.

The setup: A portfolio_companies object with fields for stage, sector, founding team, last touchpoint date, current ask, health status, and follow-up needed. A view sorted by "last touched" to surface companies that haven't been checked in with in over 30 days.

The workflow: Quarterly review prep used to take a full day. Now: ask the agent "Prepare a portfolio review summary: current status of each company, last interaction date, any that are at risk or need immediate support". The agent pulls from DuckDB and generates a structured briefing document.

What they said: "I manage 80 companies essentially alone. DenchClaw didn't just organize the data — it gave me a thinking partner. I can ask 'which of my portfolio companies are in fintech and hiring in NYC?' and I get a real answer instead of a spreadsheet I have to manually filter."

What These Stories Have in Common#

Reading through the use cases, a pattern emerges. People come to DenchClaw because:

1. They're frustrated with SaaS CRMs that cost too much, lock up their data, and require too much manual work
2. They need a system that learns their context — not a form to fill out after every meeting
3. They want AI that's actually integrated with their data, not a chatbot bolted on the side
4. They care about ownership — their contacts, notes, and history should be theirs, forever

The variety of use cases also reinforces something we believe deeply: a good CRM is really just a good database with a great interface and a smart assistant. The schema should fit the workflow, not the other way around.

Share Your Story#

If you're using DenchClaw and have a workflow that works, we want to hear about it. Join the Discord (discord.com/invite/clawd) and share in #showcase, or tag us on X. The best community stories make it into future spotlights — and sometimes directly influence which features we build next.

Your use case might be the one that helps someone else finally get organized.

Frequently Asked Questions#

How do I get started with DenchClaw for my specific use case?#

Install with npx denchclaw and tell the agent what you're trying to manage. Something like: "I'm a consultant with 15 clients. Help me set up a CRM to track relationships, projects, and notes." The agent will ask clarifying questions and scaffold the objects and fields for you. Check out what DenchClaw is for the full overview.

Is DenchClaw only for technical users?#

No. Several users in this spotlight have no technical background. The setup is command-line (one npx command), but after that, everything is conversational. You talk to the agent; it manages the database. Most non-technical users find the natural language interface more intuitive than traditional CRM UIs.

Can teams share a DenchClaw instance?#

Currently DenchClaw is primarily optimized for individual and small-team use. The multi-user collaboration features (DenchClaw Cloud with shared workspaces) are on the roadmap for 2026. In the meantime, some small teams share a single instance running on a shared server, with team members connecting via Telegram or WhatsApp.

What happens to my data if I stop using DenchClaw?#

Your data is in a DuckDB file on your filesystem and markdown files in your workspace directory. You can export to CSV at any time, open the .duckdb file with any DuckDB client, or version-control the entire workspace with git. There's no lock-in — your data is always accessible, even if you stop using DenchClaw entirely.

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