DenchClaw Team Workflow: Collaborate Without the Cloud
DenchClaw team workflow guide: collaborate on shared CRM data without cloud dependencies. Local network setup, shared objects, and team conventions.
DenchClaw Team Workflow: Collaborate Without the Cloud
"Local-first" doesn't mean solo. DenchClaw supports team workflows where multiple people share a database, collaborate on deals, and coordinate through AI — without giving data to a third-party cloud.
This guide covers how to set up effective team collaboration with DenchClaw.
The Local-First Team Model#
In a traditional SaaS CRM, team collaboration means everyone connects to the same vendor's cloud. Your data lives there, their servers process it, and you pay per seat.
DenchClaw's model: one DenchClaw instance runs on a machine (or VPS) your team controls. Team members connect to it via:
- Web UI on a shared local network
- Personal Telegram/WhatsApp bots connected to the same instance
- VPN access if remote
The database is on your hardware. Access is via your network. No monthly seat fees.
Deployment Options for Teams#
Option A: Local Network Server (Office or Colocated Team)#
One machine on your network runs DenchClaw:
# On the server machine
npx denchclaw --host 0.0.0.0Team members access http://192.168.1.X:3100 on the same WiFi. For a nicer URL, add a local DNS entry: crm.company.local.
Best for: Small teams (2-8 people) in the same physical location.
Option B: VPS (Remote or Distributed Team)#
# On a DigitalOcean/Linode VPS
npx denchclaw --host 0.0.0.0 --port 3100
# Add Nginx + Let's Encrypt for HTTPS
# Result: https://crm.yourcompany.comTeam members access via browser from anywhere. Each person's Telegram/WhatsApp bot connects to the same gateway.
Best for: Remote-first teams, companies with data residency requirements.
Option C: Dench Cloud#
Managed hosting at dench.com. Same DenchClaw stack, zero infrastructure management.
Setting Up Team Channels#
Each team member gets their own messaging channel connected to the shared instance.
For Telegram (recommended): Each person creates their own bot via @BotFather:
/newbot→ name the bot[YourName]-DenchClaw-Bot- Get the HTTP token
- Add to DenchClaw:
Gateway → Channels → Telegram → Add [person]'s bot
Now Alice uses Alice's bot. Bob uses Bob's bot. Same data, separate conversation contexts.
For Discord: Create a team server. Each team member gets a private channel where the DenchClaw bot responds to them. Shared channels for reports and alerts.
For Slack:
Configure the DenchClaw Slack app and invite it to relevant channels. Team-wide reports go to #crm-updates. Individual queries via DMs to the bot.
Owner-Based Data Organization#
Without user authentication (coming soon), track ownership via data conventions:
Owner field on every object:
Add an Owner field (text) to People, Companies, Deals, and Leads if it doesn't exist
Per-owner views:
Create views for each team member:
- "Alice's Deals" — Deals where Owner = "Alice"
- "Bob's Leads" — Leads where Owner = "Bob"
- "Team Pipeline" — all active deals, all owners
Assignment via chat:
Assign the Acme Corp deal to Bob
DenchClaw updates the Owner field and optionally notifies Bob via his channel.
Collaboration Patterns That Work#
Pattern 1: The Morning Standup Brief#
Each morning, each team member asks:
What's my day look like? Show me: my tasks due today, deals that need action, any new leads assigned to me, and anything urgent from last night.
Personalized, private — but all pulling from the same database.
Pattern 2: The Weekly Team Review#
A Friday report goes to the team channel:
Every Friday at 4pm, generate a team sales review and post to #sales:
- Pipeline by owner (bar chart would be nice)
- Deals that moved this week
- New leads by source
- Any deals at risk
- Top win of the week
Everyone sees the same data, shared context.
Pattern 3: Deal Handoffs#
When Alice hands a deal to Bob:
Handoff Acme Corp deal from Alice to Bob. Add a handoff note summarizing the deal context. Notify Bob.
DenchClaw updates ownership, logs the handoff note, and messages Bob via his channel.
Pattern 4: Collaborative Deal Rooms#
Each deal's entry document is the shared notes space:
Add to the Acme Corp deal room: engineering team wants to see a security audit before they sign. Legal review starts next week. @Bob to follow up on security docs by Thursday.
Any team member with access can read and update the deal document.
Pattern 5: Shared Dashboards#
Build a team dashboard app that everyone uses:
Build a team sales dashboard showing pipeline by owner, deals closing this week, and team win rate for the month. Make it accessible at /apps/team-dashboard.
The app runs on the shared DenchClaw instance. Everyone accesses the same URL.
Setting Team Conventions#
Document your data conventions to keep the shared database clean:
Create a document called "Team CRM Conventions" with these rules:
1. Every lead, contact, and deal must have an Owner
2. Stage changes require a note explaining why
3. All active deals need a Next Action Date within 14 days
4. Dead deals get marked Closed Lost (not left in Negotiation forever)
5. Enrich new contacts within 24 hours of adding
6. Weekly pipeline review every Monday
Post it in your team's Slack or Discord. New team members read it before they start adding data.
Access Control#
DenchClaw's current model is trust-based — all authenticated users see all data. For teams that need separation:
Separate objects per team: Keep customer success in a cs_ prefixed object set, sales in standard objects.
Read-only reporting users: Create a view-only dashboard app for stakeholders who shouldn't edit records.
Audit trails: Every AI operation is logged. Review ~/.openclaw-dench/workspace/logs/ for a full audit trail of who changed what.
Coming soon: Role-based access control with per-user permissions.
Onboarding New Team Members#
When someone joins:
- Give them access to the DenchClaw URL (or VPN access)
- Set up their personal Telegram bot and connect it
- Create their personal views:
[Name]'s Deals,[Name]'s Leads - Send them the Team CRM Conventions document
- Start with:
Welcome to our CRM. Ask me anything about how to use it.
The AI guides new users through basic operations. No training session required.
Performance for Teams#
DuckDB handles concurrent reads very well. For concurrent writes:
- Multiple people updating different records simultaneously: fine
- Multiple people updating the same record simultaneously: last write wins
- Large bulk operations: run during off-hours via scheduled cron
For teams of 5+, run DenchClaw on hardware with at least 8GB RAM for smooth performance.
For the full DenchClaw platform overview, see what is DenchClaw. For the complete team setup walkthrough, see team CRM setup guide.
Frequently Asked Questions#
Can team members edit data from their phones?#
Yes — via their personal Telegram/WhatsApp bots or the mobile browser. Full read/write access from mobile.
How do we handle someone leaving the team?#
Reassign their records: Reassign all of [person]'s open leads and deals to [new owner]. Then remove their channel connection from DenchClaw settings.
Can we have team members in different time zones collaborate?#
Yes. The async nature of messaging channels works well across time zones. The shared database always has the current state regardless of who last updated it.
Is there a way to see who made a specific change?#
The changelog in entry documents tracks AI operations with timestamps. For direct database changes, DuckDB doesn't have built-in change tracking — but the AI operation logs show which channel/session triggered each change.
How many team members can share one DenchClaw instance?#
Teams of 2-20 work well on a single instance. Beyond that, you may want to consider Dench Cloud or a more powerful server.
Ready to try DenchClaw? Install in one command: npx denchclaw. Full setup guide →
