Skip to main content

Built for humans.
Operable by agents.

TappHQ is the first nonprofit operations platform designed to be driven by AI agents - not just assisted by them. AI lives inside every module. Your custom agents run on AgentHQ. And any external AI agent can operate the platform natively through open protocols.

See also: the cross-module intelligence layer

Start Free Developer Docs
228
Platform Tools
11
HQ Brands
4
Agent Protocols
3
Agentic Layers
Three-Layer Architecture

AI works here. Your agents work here.
Any agent works here.

Most platforms add AI as a feature. TappHQ treats agent operability as a first-class product surface - at every layer of the stack.

Layer 1: AI Works Here

AI is embedded in every module. Content generation, grant writing, compliance analysis, predictive scoring, meeting intelligence, executive briefings. Your team uses AI without thinking about it.

  • Ask TappIQ assistant in every context
  • Grant matching and draft generation
  • Compliance risk scoring and alerts
  • Report Builder and Content Studio

Layer 2: Your Agents Work Here

Build custom agents that operate inside TappHQ using the AgentHQ builder, 80+ templates, or Workflow Automation. Your agents have full context about your org and run with cost ceilings and approval gates.

  • AgentHQ visual builder
  • 80+ production-ready agent templates
  • Workflow Automation and triggers
  • Cost tracking and human-in-the-loop

Layer 3: Any Agent Works Here

External AI agents operate TappHQ natively. The platform publishes its capabilities through open standards - MCP, A2A, webhooks, and a REST API. Claude, ChatGPT, Cursor, and any compatible agent can drive the platform.

  • MCP Server for direct tool access
  • A2A protocol for agent-to-agent delegation
  • Capabilities discovery endpoint
  • Scoped API keys with tier-based access
Agentic Surface

Five ways agents operate the platform

TappHQ exposes its operations through open, standard protocols so agents don't need custom-built integrations.

MCP Server

A Model Context Protocol (MCP) server lets Claude Desktop, Claude Code, Cursor, and any MCP-compatible client connect and use all 228 platform tools directly. Pro plans: read-only. Enterprise: full read and write.

A2A Protocol

Agent-to-Agent delegation. External agents send secure task requests; TappHQ agents accept them, do the work, and return results. This lets agents from different systems work together - your compliance agent can hand work off to an external research agent.

Capabilities Discovery

TappHQ publishes a public directory of what the platform can do. Agents discover the available tools and their access rules on their own, with no setup required - the foundation for routing work to the right agent automatically.

REST API + Webhooks

Full REST API with OpenAPI 3.1 specs, interactive reference, and 13 scoped permission groups. Webhooks notify external agents when platform state changes - deadlines approach, risks surface, grants match.

Scoped API Keys

Org-scoped keys with tier-aware permissions. Free tier gets public capabilities; Pro unlocks read MCP; Enterprise gets read-write MCP plus A2A. Rotate, revoke, and audit from the dashboard.

Audit + Rate Limits

Every agent call is logged to the audit log with the calling agent, tool, cost, and result. Rate limits and credit ceilings prevent runaway agent spend before it happens.

Real Scenarios

What agent-operable actually looks like

Claude Code reviews your workflows

A development team uses Claude Code with the TappHQ MCP server connected. Claude Code queries workflows, suggests optimizations, and drafts changes - all from the IDE. No context switching, no copy-paste.

ChatGPT acts as your ED assistant

Your Executive Director connects TappHQ to ChatGPT via the MCP connector. They ask "What's my top risk this week?" and ChatGPT reads your active alerts and drafts a board update.

A2A delegation across agents

Your grant-writing agent needs funder research. It sends an A2A task to an external research agent, which returns structured funder intelligence. Your agent incorporates it into the draft and submits for review.

Cursor bulk-updates compliance items

An ops manager needs to reclassify 40 compliance items. Cursor with the TappHQ MCP server connected reads the current state, proposes the reclassification, and applies it after approval - in under a minute.

Tier Access

Agent capabilities by plan

Free & Starter

  • Public capabilities endpoint
  • REST API (read-only)
  • Webhooks
  • Embedded AI features in-app

Pro

  • Everything in Starter
  • MCP Server (read)
  • Scoped API keys
  • Agent audit log
  • Custom AgentHQ agents

Enterprise

  • Everything in Pro
  • MCP Server (read + write)
  • A2A protocol
  • Priority agent rate limits
  • Dedicated agent support
TAPPIQ Memory

Personalized to your organization. Automatically.

What makes an agentic platform truly useful isn't just that agents can run - it's that they know who they're working for. TAPPIQ Memory is the personalization layer that makes “AI works here” specifically valuable for your organization. All 6 dimensions - preferences, behavioral patterns, decision history, voice, relationships, and contradiction detection - are applied automatically to everything the AI does across the platform.

The longer your organization uses TappHQ, the more institutional knowledge accumulates. Agents that run six months from now will know your organization far better than agents running today - without you doing anything extra. That accumulated context is what no competitor can replicate overnight.

Manage memory at /settings/memory. Starter+ required.

Evaluating from a technical angle?

See the architecture detail, security model, integration decision tree, and full tool inventory at /features/agent-platform-technical.

Technical evaluator page

Ready to run your nonprofit with agents?

Start free and connect your first agent in minutes. Upgrade when you're ready to let external agents operate the platform.

Start FreeRead Developer Docs