OpenClaw is an open-source AI executive assistant that connects to your tools (email, calendar, Slack, etc.) and automates workflows. It's built for teams who want AI that actually does work, not just answers questions.

Do I need technical knowledge?

Not at all. That's the whole point of SetupOpenClaw — we handle every technical detail. You just tell us what you want your AI assistant to do.

Can you set this up remotely?

Yes. Our Hosted Setup and Mac Mini Remote tiers are fully remote. We can deploy for teams anywhere in the UK (and worldwide). Only the Mac Mini In-Person tier requires Leeds & West Yorkshire proximity.

Is this related to the OpenAI acquisition?

No. OpenClaw is an independent open-source project. It uses various AI models and is not owned by or affiliated with OpenAI.

Can I set up OpenClaw myself?

Absolutely — it's open source. But most teams find that security hardening, integration configuration, and ongoing maintenance take significantly more time than expected. We do it in hours with enterprise-grade security.

What if something goes wrong?

During the 14-day hypercare period, we actively monitor and fix any issues. After that, we offer support packages. Plus, every installation includes audit trails and monitoring.

What infrastructure do I need?

For Hosted Setup: nothing — we deploy to a cloud VPS (~£5-10/mo). For Mac Mini tiers: a Mac Mini (~£500). We handle all software configuration.

Can I have multiple agents?

Yes. Each additional agent is £350. Teams often run separate agents for different departments or workflows.

Yes. We use Docker sandboxing, Composio middleware, OAuth token management, audit trails, and read-only defaults. Your data stays on your infrastructure — we never store it.

What about ongoing maintenance?

OpenClaw requires minimal maintenance after setup. We provide documentation for common tasks. For hands-off operation, ask about our maintenance packages.

What is the most important file in an OpenClaw workspace?

SOUL.md is the foundational file. It defines the agent’s core identity, values, communication style, and behavioural boundaries. Every other workspace file builds on the personality and constraints established in SOUL.md. Without it, the agent has no consistent character or guardrails.

How long does it take to configure all 8 workspace files?

A complete workspace configuration typically takes 4–8 hours for a DIY setup, covering SOUL.md, IDENTITY.md, AGENTS.md, USER.md, TOOLS.md, MEMORY.md, HEARTBEAT.md, and BOOT.md. With SetupClaw’s managed service, configuration is completed during the same-day deployment as part of the setup package.

Can I configure workspace files myself without SetupClaw?

Yes — all workspace files are standard markdown that you can edit in any text editor. Our documentation covers every field across all 8 files. SetupClaw’s managed service is for teams who want expert configuration, security hardening, and 14-day hypercare included.

OpenClaw HEARTBEAT.md Configuration: Scheduled Autonomous Tasks

What This File Controls

HEARTBEAT.md transforms your agent from a passive chatbot into an active participant. A chatbot waits for messages; an agent with HEARTBEAT.md acts on a schedule — monitoring data sources, sending notifications when conditions are met, and generating reports at defined intervals. This is the file that makes the difference between an assistant and an autonomous operator.

The runtime reads HEARTBEAT.md to configure three types of scheduled behaviour: monitoring (high-frequency checks, typically every 15–30 minutes), notifications (condition-triggered alerts, typically every 1–2 hours), and reports (scheduled summaries, typically daily or weekly). Each type operates independently and can be configured with different schedules.

HEARTBEAT.md requires the agent to be running continuously — it only works on deployments with persistent process hosting (e.g., a VPS). If you are running OpenClaw only in response to manual triggers, HEARTBEAT.md schedules will never fire. All SetupClaw deployments include VPS hosting with continuous process management from the £770 tier.

Why it matters: A chatbot waits. An agent acts. HEARTBEAT.md is the difference.

Field-by-Field Breakdown

3 fields that shape your agent

heartbeatMonitorOptional

What the agent should proactively watch for.

heartbeatMonitor: "e.g. Check Slack #alerts every 15 minutes for new issues..."

heartbeatNotifyOptional

Conditions that trigger proactive notifications.

heartbeatNotify: "e.g. DM the user if a P1 ticket is unassigned for 30+ minutes..."

heartbeatReportOptional

Scheduled reports the agent generates automatically.

heartbeatReport: "e.g. Send a daily summary of resolved tickets at 17:00..."

Real-World Examples

FinleyManchester · 8 people · Accounting

## heartbeatMonitor
- Every 30 minutes: check GOV.UK RSS feed for HMRC guidance updates relevant to current client matters
- Every 30 minutes: check shared calendar for filing deadlines within the next 14 days

## heartbeatNotify
- If a filing deadline is within 7 days and no completion note exists: DM the responsible associate
- If HMRC publishes updated guidance on a topic with an active client matter: post to #tax-queries
- If a partner has not responded to a flagged research query within 4 hours: send a reminder

## heartbeatReport
- Monday 08:00 GMT: weekly filing deadline summary for all clients
- 1st of each month: monthly compliance status report
- Quarterly: prepare Board reporting pack data on research utilisation

Common Mistakes

Configuring without continuous deployment

HEARTBEAT.md schedules only fire when the agent process is running continuously. If you run OpenClaw only on-demand (e.g., triggered by messages), heartbeat tasks never execute. Ensure your deployment includes persistent hosting.

Over-scheduling notifications

An agent that sends 20 notifications a day creates alert fatigue. Your team starts ignoring alerts, defeating the purpose. Batch low-priority notifications and use cooldown periods between alerts on the same topic.

Heartbeat tasks requiring capabilities not in TOOLS.md

If HEARTBEAT.md schedules a check that requires an API not listed in TOOLS.md, the task fails silently. Audit every heartbeat task against TOOLS.md capabilities.

Report timing misaligned with team working patterns

A daily report delivered at 06:00 GMT is useless if your team starts at 09:00. A Monday briefing sent at 17:00 misses the planning window. Align report schedules with when the team actually reads them.

How SetupClaw Handles This

All SetupClaw deployments run on VPS hosting with continuous process management, meaning heartbeat schedules work from day one. Our specialist configures monitoring, notifications, and reports based on your team’s actual working patterns, including UK bank holiday handling and GMT/BST timezone awareness.

£770

VPS deployment with continuous hosting + heartbeat schedules configured to your team’s timezone + cross-file validation against TOOLS.md

£1,100

All above + alert fatigue prevention (batching, cooldowns, urgency categorisation) + bank holiday awareness

£2,200

All above + custom monitoring integrations + quarterly heartbeat optimisation review

All 3 fields configured with appropriate schedules
Schedules aligned to team working hours and timezone
UK bank holiday handling configured
Cross-validated against TOOLS.md capabilities
Alert fatigue prevention rules implemented

Configured during deployment (day 1–2). Alert tuning during first week (Standard+).

Advanced Topics

GMT/BST timezone handling

UK-based agents need to handle the GMT/BST clock change. There are three approaches: fixed UTC (simplest but report times shift for users), dual schedule (separate winter/summer configs), and timezone-aware cron (runtime resolves Europe/London automatically). For most UK businesses, timezone-aware cron is recommended — it requires no manual updates when the clocks change.

## heartbeatReport
# Timezone-aware approach (recommended for UK)
- Monday 09:00 Europe/London: weekly briefing
- Daily 17:30 Europe/London: end-of-day summary
# These automatically adjust for GMT/BST changes

Example Configurations

Persona A — MeridianChief of Staff

# HEARTBEAT.md -- Scheduled Autonomous Tasks

## heartbeatMonitor
- Every 15 minutes: check for urgent Slack messages from VPs
- Every 30 minutes: scan calendar for conflicts within the next 48 hours

## heartbeatNotify
- If an action item is overdue by >24 hours: DM the principal with context
- If a board-related email has been unresponded for >4 hours: flag immediately

## heartbeatReport
- Monday 08:00 Europe/London: priorities briefing for the week
- Friday 17:00 Europe/London: week-in-review summary
- Monthly (1st): stakeholder map update and relationship health check

Persona B — Axiom (Veteran Growth Hacker)

# HEARTBEAT.md -- Scheduled Autonomous Tasks

## heartbeatMonitor
- Every 15 minutes: check funnel metrics for >15% deviation from baseline
- Every 30 minutes: monitor ad spend pacing against daily budget

## heartbeatNotify
- If an experiment shows early statistical significance: alert immediately
- If ad spend pacing exceeds 120% of daily target: pause and notify

## heartbeatReport
- Monday 09:00 Europe/London: weekly growth digest (funnel metrics, active experiments, top findings)
- Monthly: growth playbook update with winning/losing experiment summary

Persona C — Lumen (Scientific Researcher)

# HEARTBEAT.md -- Scheduled Autonomous Tasks

## heartbeatMonitor
- Every 30 minutes: check PubMed RSS feeds for new publications in assigned domains
- Every 30 minutes: check bioRxiv preprint server for relevant preprints

## heartbeatNotify
- If a high-impact journal publishes in a tracked domain: alert within 1 hour
- Weekly: publication digest of all new relevant papers

## heartbeatReport
- Monthly: domain knowledge map update (consensus positions, evidence gaps, emerging themes)
- Quarterly: comprehensive landscape report for grant review preparation

Persona D — Narrative (Senior Content Creator)

# HEARTBEAT.md -- Scheduled Autonomous Tasks

## heartbeatMonitor
- Every 15 minutes: monitor traffic for spikes >200% of daily average
- Every 30 minutes: check editorial calendar for gaps within the next 14 days

## heartbeatNotify
- At 48 hours post-publication: content performance snapshot (views, engagement, conversions)
- If a competitor publishes on a tracked topic: alert with link and brief analysis

## heartbeatReport
- Monday 09:00 Europe/London: weekly content digest (published, scheduled, performance highlights)
- Monthly: full content audit (top performers, underperformers, gap analysis)

Persona E — Architect (Chief Engineer)

# HEARTBEAT.md -- Scheduled Autonomous Tasks

## heartbeatMonitor
- Every 15 minutes: check latency and error rate against SLA thresholds
- Every 30 minutes: check for PRs open >48 hours without review

## heartbeatNotify
- If SLA breach detected (uptime <99.95% or p99 >200ms): alert immediately
- If PagerDuty fires an active alert: surface context and suggested actions
- If a service restarts unexpectedly: log and notify on-call engineer

## heartbeatReport
- Monday 09:00 Europe/London: engineering health dashboard (uptime, latency, incidents, PR velocity)
- Monthly: ADR index update and tech debt inventory

HEARTBEAT.md — Scheduled autonomous tasks