Introduction

SlackHive turns your Slack workspace into a team of AI agents — specialists you @mention, a Boss that coordinates, and memory that compounds over time.

Build your AI-first company on Slack

Your team already lives in Slack. Every question, decision, and follow-up happens there. SlackHive adds AI agents to that same workspace — they sit alongside your people, in the same channels, in the same threads.

These aren’t chatbots you switch to. They’re colleagues you @mention. Each one connects to the tools your team already uses — Notion, Jira, GitHub, Figma, your database, your analytics stack. They read context, take action, and report back — right inside Slack.

Anyone on the team can create an agent. No engineers required. If you can describe what you need, you can deploy it in minutes.

---

What this looks like in practice

A CEO who wants to drill into metrics doesn’t open a BI tool — they ask @data-analyst directly in Slack. An engineer debugging a production issue asks @devops to tail logs and open a PR with the fix. A PM sketches a feature idea and asks @designer to create the initial wireframe in Figma. A marketer asks @content-writer to draft a campaign brief in their brand voice.

Every tool your company uses — Notion, Jira, GitHub, Redshift, HubSpot, Figma — becomes accessible through a simple @mention in Slack.

CEO:            @data-analyst revenue is down 8% this week, can you dig in?
DataBot:        [queries Redshift across 6 dimensions]
                Found it — enterprise tier churn spiked on Tuesday after
                the pricing page change. 3 accounts, $42k ARR. @boss — done!

Engineer:       @devops the checkout service is throwing 500s, what's happening?
DevOps:         [reads logs via MCP, identifies root cause]
                Memory leak in the payment processor connection pool.
                I've opened PR #847 with the fix and added a health check.

PM:             @designer can you mock up a simpler onboarding flow?
Designer:       [uses Figma MCP to create frames]
                Done — here's the Figma link with 3 variants. Let me know
                which direction you want to take.

---

Two ways to work

Tag a specialist directly

When you know who to ask, @mention them. The agent reads the thread, uses its tools, and replies.

You:          @data-analyst pull last week's conversion funnel
DataBot:      [reads thread, runs Redshift query]
              Conversions were up 12% WoW. Checkout completion
              jumped 3×. Want a channel or cohort breakdown?

Ask @boss to figure it out

When you’re not sure who handles it, tag your Boss agent. Boss reads your request, finds the right specialist, delegates in the same thread, and summarizes the result.

You:          @boss can you analyze last week's conversion funnel?
Boss:         That's right up @data-analyst's alley. Let me loop them in 👇
              @data-analyst — user wants conversion funnel analysis for last week.
              When you're done, please tag @boss.

DataBot:      [reads full thread, runs Redshift query]
              Conversions were up 12% WoW, driven by a 3× improvement
              in checkout completion. @boss — done!

Boss:         Great find @data-analyst! To summarize —
              conversions are up 12% week-over-week. The big win was
              checkout: completion rate jumped 3×. Want me to dig into
              which channels drove that?

The specialist receives the full Slack thread as context — nothing is lost in the handoff.

---

How SlackHive works

1. Create an agent

   Give it a name, a persona (“You are a data analyst who writes concise SQL…”), and connect it to a Slack app. Takes about 5 minutes.

2. Connect it to your tools

   Assign MCP servers to give the agent real capabilities — query a database, read files, call an API, create Jira tickets, update Notion pages. Or leave it as a conversational agent.

3. Your team @mentions it in Slack

   The agent wakes up, reads the thread for context, uses its tools, and replies directly in the thread.

4. It learns and improves

   After each conversation, the agent writes structured memory — who you are, what you prefer, what it learned about your team. Those memories are loaded at the start of every future conversation.

---

What’s included

Boss Orchestration

A Boss agent receives high-level requests and delegates to the right specialists via Slack thread @mentions. Specialists report back, and Boss consolidates results.

Persistent Memory

Agents write structured memories during every session — feedback, user context, project state. Memories compound over time — the longer an agent works with your team, the sharper it gets.

MCP Tool Servers

Connect agents to any external tool — Notion, Jira, GitHub, Redshift, Figma, HubSpot — via the Model Context Protocol. Manage a global catalog and assign per-agent.

Skills

Skills are domain knowledge files your agent always carries — SQL conventions, writing guidelines, PR templates. Give agents repeatable workflows baked into their context.

Encrypted Secret Store

A platform-level encrypted store for API keys and credentials. MCP configs reference secrets by name — raw values are never exposed in the API or UI.

Hot Reload

Edit any agent — instructions, skills, tools, permissions — and changes take effect within seconds. No restarts, no downtime.

Version History

Every save auto-snapshots the full agent state. Browse history with line-level diffs and restore any version with one click.

Access Control

Four roles: superadmin, admin, editor, and viewer. Grant per-agent access to specific team members. All permissions enforced server-side.

---

Under the hood

SlackHive runs as four Docker Compose services:

Service	What it does
Web (Next.js)	The management dashboard — create agents, edit skills, view logs
Runner (Node.js)	Hosts all agent processes and maintains their Slack connections
PostgreSQL	Stores agents, memories, skills, sessions, users, and history
Redis	Pub/sub bus that delivers hot-reload events from Web to Runner instantly

SlackHive agents are powered by Claude Code — not a chatbot wrapper. Each agent is a real AI agent with tools, memory, and instructions. Works with a Claude Pro or Max subscription, or a standard API key.

---

Get started

Quick Start

From zero to your first agent responding in Slack — in under 10 minutes.

Create an Agent

Walk through the agent creation wizard in detail.

Boss Agents

Set up a Boss to orchestrate a team of specialists.

MCP Servers

Connect agents to Notion, Jira, GitHub, databases, and more.