Organizations
Flowyble is multi-tenant. Every user belongs to an organization, and all data (agents, tools, runs, credits) is scoped to that organization. Team members within an organization share the same agents, tools, and credit balance.Agents
An Agent is an autonomous AI entity that receives a user message, reasons about it, optionally calls tools, and produces a final output. Agents are configured with:| Setting | Description |
|---|---|
| Behavior Prompt | System prompt that defines the agent’s personality and instructions |
| LLM Model | Which language model powers the agent (GPT-4o, Claude Sonnet, etc.) |
| Max Steps | Maximum number of tool-call iterations before the agent stops (1–12) |
| Conversation Memory | When enabled, the agent remembers previous messages in a conversation |
| Knowledge Base | Optional document collection the agent can search |
| Tools | Functions the agent can call during execution |
Tools
A Tool is a multi-step workflow automation. Unlike agents, tools follow a predefined sequence of steps — each step calls a built-in function or integration with specific parameters. Tools are deterministic and repeatable.Runs
Every time an agent or tool executes, it creates a Run record:- AgentRun — tracks status, tokens, credits, steps, quality score, and final output
- ToolRun — tracks status, input/output data, and execution metrics
Conversations
A Conversation groups multiple agent runs into a multi-turn dialogue. When conversation memory is enabled, the agent receives the full conversation history with each new message, enabling context-aware responses.Credits
Flowyble uses a credit-based billing system. Each LLM call consumes credits based on:- Input tokens — tokens sent to the model
- Output tokens — tokens generated by the model
- Credit rate — per-model pricing (varies by provider and model size)
Integrations
Integrations connect Flowyble to external services (e.g. Google, Microsoft, Slack). Each integration can provide:- OAuth2 authentication — secure token-based access
- Integration tools — pre-built functions that call external APIs
- Integration accounts — stored credentials per organization
Webhooks
Webhook triggers allow external systems to start agent or tool runs by sending HTTP requests to a unique webhook URL. Webhooks support optional HMAC signature verification for security.System Agents
System agents are platform-level AI agents that run automatically during execution:| Type | Purpose |
|---|---|
| Chat | Powers the built-in chat interface |
| ToolSelector | Dynamically selects which tools an agent should use |
| QualityChecker | Evaluates agent output quality (pass/warn/fail) |
| ErrorAnalyzer | Analyzes execution errors and suggests fixes |