Framework · Conversational Bots

Parlant

Parlant is a framework for building conversational agents whose behaviour is controlled by condition-action guidelines that the engine matches and applies through structured Attentive Reasoning Queries.

Description

Parlant is a Python framework for building customer-facing conversational agents. Behaviour is defined as granular condition-action guidelines rather than free-form system-prompt instructions, and an engine pipeline matches only the relevant guidelines for the current conversational state, calls associated tools, and composes the message. The message composer uses Attentive Reasoning Queries, structured reasoning blueprints, to keep the agent adhering to its matched guidelines. It is released under Apache 2.0.

Solution

On each turn, the engine runs a pipeline: a GuidelineMatcher filters the full set of condition-action guidelines down to those relevant for the current conversational state, a ToolCaller executes the tools associated with matched guidelines, and a MessageComposer generates the response using the matched guidelines, tool results, and history under ARQ-based structured reasoning. The pipeline iterates until reaching a stable state.

Primary use cases

  • building controlled customer-facing chat agents
  • enforcing behavioral guidelines in conversation
  • tool-using conversational workflows
  • explainable agent decisions

Open the full interactive page

Diagram, neighbourhood map, code examples, related patterns and full provenance.