AI Automation for Team Slack Workspaces
Pulse: Claude in your team's Slack, installed on your workspace, owned by you
Productized install for agencies and SMBs. Six Claude commands firing on schedule. $3,500 install, five days, no monthly platform fee.
Live map
activeSystem type
overlayReplacement
not requiredNext step
scanWhy your team's Slack feels like noise instead of a system
Your team's Slack runs hot all day. Notifications stack, threads scatter, the same questions get asked four times a week, and the institutional knowledge that should compound is buried under message-volume. Most teams react to this with more channels, more keywords, more notification preferences. The structural problem is that Slack itself has no operating intelligence. It moves messages; it does not produce summaries, surface anomalies, or run scheduled work on your behalf.
The fix is not another Slack app from the marketplace. The Slack app store is full of single-purpose integrations that solve a tile of the problem and never compose. The fix is a real AI coworker living inside your workspace with persistent memory, scheduled commands, and the ability to do actual work on demand. That is what Pulse is.
What Pulse actually is
Pulse is a productized install of Claude Code configured to run inside a Slack workspace your team already uses. The install puts a Docker container on a VPS you control. The container holds the Slack bot credentials, the Anthropic API key, and a directory of commands written in plain Markdown. It listens for @-mentions, fires scheduled commands via cron, and posts results back to the Slack channel you configure per command.
Adding a command after install is a Markdown file plus a config entry plus a container restart. The starter pack ships six commands so your team has a working system on day one. Most teams add their seventh and eighth command within the first month of use, because the muscle memory of "if it is repetitive, write a command" kicks in fast once the infrastructure is there.
The six commands that ship at install
- standup-pulse. Daily morning digest posted to a configured channel. It reads from your shared sources (a Google Sheet, a Notion database, a CRM via MCP) and summarizes what shipped yesterday, what blocked, and what is up today. Replaces the 15-minute manual standup write-up your ops lead does at 8 AM.
- eod-recap. End-of-day rollup of activity, blockers, and tomorrow's focus. Pulls from the same shared sources as standup-pulse but reads the close-of-day state and surfaces what is incomplete.
- weekly-digest. Friday afternoon summary of the week's wins, decisions made, and metrics that moved. Posts to a leadership or all-hands channel. The artifact your founder used to write manually on Sunday night.
- inbox-triage. Scans the inboxes you connect (shared team inbox, founder inbox, support queue) and surfaces what genuinely needs human attention versus what can wait or auto-respond. Hands you a ranked list, you decide what to action.
- research. @-mention Pulse with a topic ("@Pulse research how RecruitCRM handles webhook auth") and Pulse returns a sourced briefing in-thread within a couple of minutes. Backed by Claude plus your configured tool list.
- ask.Generic Q&A. @-mention Pulse in any channel with a question ("@Pulse who owns the Q3 launch?", "@Pulse what's the status on the Bear account?") and it answers from your workspace context. The default fallback when nothing else fits.
How Pulse is built
Pulse runs as a single Docker container on a Linux VPS. Inside the container is a Node.js server (server.mjs) that handles Slack events plus a cron-driven heartbeat (heartbeat.mjs) that fires scheduled commands. Each command is a Markdown file with a system prompt plus optional context (data sources, tool list, output format). When a command runs, the server invokes claude -p with the right context and posts the response back to the configured channel.
Per-channel workspaces let you give different teams different defaults. Hooks let you gate Claude's actions on your own logic: require human approval before a command sends to a client-facing channel, route certain outputs to a review queue first, log every command invocation to an audit table. The template repository is at github.com/crashlattice57/pulse-template (v0.1.0 shipped 2026-04-26, 12 passing tests).
The five-day install
Day 1: VPS provisioning. We stand up a VPS (or work against one you already run), install Docker, and clone the template repo. By end of day one, the container is running and responding to a sandbox Slack channel under your workspace.
Day 2-3: Slack app creation, command tuning. You create the Slack app from a manifest we provide; we configure the six starter commands against your team's actual data sources (your shared sheets, your CRM, your inbox). Each command is tested end-to-end before we move on.
Day 4: Team working session. A 60-90 minute live session where every person on the team uses Pulse on real work. The session surfaces the rough edges that only appear in actual usage. We fix the top three same-day.
Day 5: Handoff. You receive VPS root credentials, the Slack app under your admin, the Anthropic API key under your billing, and a Runbook covering the most common ops tasks (rotate keys, add a command, edit a prompt, view logs, restart the container).
Pricing: Standard vs Premium
Two install tiers. Pick based on whether you want to extend the install yourself or have Moore IQ keep building new commands for you over time.
| Standard | Premium | |
|---|---|---|
| Install fee | $3,500 one-time | $5,000 one-time |
| Optional retainer | $500/mo | $750-$1,000/mo |
| Retainer covers | Uptime monitoring, prompt drift fixes, API change patches | Above + ongoing command development, no hour cap |
| Commands at install | Six starter commands | Six starter + two custom-built for your team |
| Best for | Teams that ship the starter pack and extend on their own | Agencies that want a build partner across multiple client installs |
What you own at handoff
Everything. The VPS root credentials are yours. The Slack app is registered under your workspace admin. The Anthropic API key is provisioned under your billing. The template repository is forked into a GitHub org you control. The six starter commands, the cron schedules, the hooks configuration, the audit log schema. All of it lives in your infrastructure and runs without Moore IQ in the loop.
If you take the retainer, we keep a scoped SSH key plus a Slack admin invite so we can do the maintenance work. If you cancel the retainer, you revoke both at the end of the 30-day notice period. The system keeps running. This is the same ownership model used in the Recruiting OS install, with smaller surface area and a tighter starter scope.
When Pulse vs Recruiting OS
Both are productized Hermes/Claude installs on infrastructure you own. Different scopes.
- Pulse is for horizontal team operations: standup, recap, digest, triage, research, ask. Six commands tuned to general team workflows. Fits any agency or SMB team that wants Claude as a coworker in their Slack.
- Recruiting OS is for recruiting firms specifically. Five skills tuned to recruiting workflows (pipeline overview, candidate research, JD formatter, BD signal scan, daily recap). Wired to your ATS and your enrichment stack. Same install shape, but recruiter-specific from day one.
If your team is a recruiting firm, start with Recruiting OS. If you are anything else, Pulse is the right entry. Some agencies run both: Pulse on their internal workspace, Recruiting OS on a recruiting-firm client's workspace.
Want to see Pulse thinking before you commit? Run the free AI Operations X-Ray. It scans your domain plus a short intake, and ranks the five highest-leverage automations you could ship next. Many of them are exactly the kind of thing Pulse runs.
Related
- Recruiting OS— the recruiter-specific productized SKU
- Claude Code for teams— the deeper architectural reference Pulse is built on
- Custom MCP server development— for teams that want to extend Pulse with their own tool access
Official resources
Frequently asked questions
- What is Pulse in one sentence?
- Pulse is a productized install that puts Claude into your team's Slack as a real coworker: it runs scheduled commands (daily standup, EOD recap, weekly digest, inbox triage), answers ad-hoc questions in a thread, and lives on a VPS you own. Five-day delivery, $3,500 install, optional $500/mo retainer.
- How is this different from just adding the Claude Slack app?
- The official Claude Slack app is a chat interface. Pulse is a full configured Claude Code system with persistent memory, scheduled cron commands, hooks, and the ability to extend with custom Markdown commands tailored to your team. It is the difference between a calculator and a coworker. You can add new commands by writing a Markdown file, restart the container, and Pulse starts running them.
- Who is Pulse for?
- Agencies running 5-30 client accounts who want a Claude-powered ops layer per client (one Pulse per workspace), and SMBs with internal teams of 5-30 people who want a few well-defined automations firing into Slack on a schedule. Not for solo operators (the team-scale leverage is the point) and not for other AI builders (you don't need this; you'd build it yourself).
- What commands ship by default?
- Six starter commands: standup-pulse (daily AM digest), eod-recap (end-of-day rollup), weekly-digest (Friday wins/decisions), inbox-triage (scans linked inboxes for what needs human attention), research (@-mention Pulse with a topic, get a sourced briefing in-thread), and ask (generic Q&A using your workspace context). Each lives at commands/<name>/command.md and is editable in plain Markdown.
- Can I add my own commands after install?
- Yes. Create a new Markdown file at commands/<name>/command.md, add an entry to pulse.config.yaml with a cron schedule and target channel, restart the container. No code changes. The starter commands give you the pattern; new commands take 15-30 minutes to add once you understand the shape.
- Where does my data live?
- On the VPS you own. Pulse is a Docker container running on a Linux VPS (typically Hetzner or DigitalOcean, 2 vCPU, 4 GB RAM, around $10-20/month). Slack tokens, your Anthropic API key, and command state all live in environment variables on that VPS. The Anthropic API does not train on commercial API traffic per their terms.
- What happens if I cancel the retainer?
- The system keeps running. Pulse is Docker plus Markdown plus a config file in a repo you own. The retainer covers uptime monitoring, prompt drift fixes, and API change patches. Without it, you maintain the system yourself or hire any TypeScript or DevOps freelancer to dig in. There is no kill switch on our end.
- Is there a difference between Standard and Premium?
- Yes. Standard is $3,500 install + $500/mo retainer. Premium is $5,000 install + $750-$1,000/mo retainer and includes ongoing command development (the team builds new commands for you on request, no per-command fee). Premium fits agencies that want to keep extending the install; Standard fits teams that want to ship the six starter commands and run them as-is.
Next step
Want this mapped for your business?
Run the 90-second AI Operations X-Ray. Free, no credit card.