Core Concepts

HeartBeatAgents is built around eight core concepts. Understanding these will help you design effective agents and get the most out of the platform.

Agents

An agent is an autonomous AI entity with a defined role, personality, and set of capabilities. Each agent has a primary LLM provider and model, standing orders that govern its behavior, access to specific tools and channels, and its own memory store. Agents operate independently. You can run multiple agents, each specialized for a different function (support, sales, engineering, etc.).

Key properties: name, description, provider, model, standing orders, autonomy level (supervised, semi-autonomous, fully autonomous), status (active, paused, archived).

Conversations

A conversation is a thread of messages between a user and an agent. Conversations maintain state across multiple turns and can span hours, days, or weeks. Each conversation belongs to a specific agent and is associated with a specific channel. Conversations are the primary unit of interaction and the source of episodic memory.

Key properties: conversation ID, agent ID, channel, participant identities, message history, status (active, resolved, escalated), metadata.

Tools

Tools are discrete capabilities that an agent can invoke during a conversation. Each tool has a name, a description (which the LLM reads to understand when to use it), input parameters, and an execution handler. Tools are provided by integrations. When you connect a Stripe integration, tools like stripe_get_customer and stripe_create_refund become available to your agents.

Key properties: tool name, description, parameter schema (JSON Schema), integration source, required permissions, rate limits.

Memory

Memory gives agents continuity across conversations. HeartBeatAgents implements three memory types:

• Episodic: Stores conversation history. Enables agents to recall prior interactions with a user or about a topic.
• Semantic: Stores extracted facts and knowledge. Enables agents to remember specific details like preferences, account information, and decisions.
• Procedural: Stores learned behavioral patterns. Enables agents to reuse successful multi-step processes for recurring task types.

All memory types use vector embeddings for retrieval, allowing agents to find relevant memories based on semantic similarity rather than keyword matching alone.

Channels

Channels are communication platforms where agents interact with users. HeartBeatAgents supports 17 channels including Slack, Discord, Microsoft Teams, Telegram, WhatsApp, SMS, email, web chat, and more. Each channel has its own adapter that normalizes messages into a standard format. An agent can be connected to multiple channels simultaneously, and conversations maintain context regardless of which channel they occur on.

Key properties: channel type, connection status, authentication credentials, message format, rate limits, supported features (threads, reactions, file attachments).

Integrations

Integrations connect your agent to external services. Each integration authenticates with a third-party API (typically via OAuth 2.0) and exposes a set of tools that agents can use. Available integrations include Google Workspace, Microsoft 365, GitHub, GitLab, Jira, Linear, HubSpot, Salesforce, Stripe, Notion, Confluence, and more. When you connect an integration, its tools are automatically discovered and made available to your agents.

Key properties: service name, authentication method, connection status, available tools, required scopes, refresh token status.

Standing Orders

Standing orders are persistent instructions that define an agent's behavior across all conversations. Unlike a single system prompt, standing orders are structured: they can include role definitions, behavioral rules, knowledge boundaries, escalation policies, and operational limits. Standing orders are versioned, so you can roll back to a previous version if a change produces undesirable behavior. They are the single most important factor in determining agent quality.

Key properties: order text, version number, created date, active status, associated agent, change history.