How victor-hugos/agente-de-atendimento-a works
ARISE is a prototype conversational booking agent for an automotive-care business. In the demonstrated terminal workflow, a customer message enters through the CLI, the agent builds context from the prompt, service data, and session, then asks OpenAI for a reply and may invoke tools to update a booking or record a human handoff. Session state remains in memory; each exchange is costed and logged. The README also describes a server and owner panel, but the tree contains no panel implementation or server conversation routes, so those links are not inferred.
Main entry points: agent.ts Conversation agent · cli.ts Terminal chat · config.ts Runtime configuration · session.ts Conversation and order state
Components 12
Entry channels
- Terminal chat
src/cli.ts - HTTP server
src/server.ts
Conversation runtime
- Conversation agent
src/agent.ts - Agent instructions
prompts/system.md - Service knowledge
data/arise.json
Booking workflow
- Agent tools
src/tools.ts - Conversation and order state
src/session.ts
Usage and operations
- Runtime configuration
src/config.ts - Costing and conversation log
src/logger.ts
Other components
- Customer
- Health client
- OpenAI API
Connections 15
- Customer → Terminal chat: enters messages
README.md - Terminal chat → Conversation agent: sends messages
src/cli.ts - Terminal chat → Conversation and order state: reads and resets
src/cli.ts - Conversation agent → Conversation and order state: gets session
src/agent.ts - Conversation agent → Agent instructions: loads instructions
src/agent.ts - Conversation agent → Service knowledge: loads service data
src/agent.ts - Conversation agent → OpenAI API: requests completions
src/agent.ts - Conversation agent → Agent tools: executes tool calls
src/agent.ts - Agent tools → Conversation and order state: updates order and handoff
src/tools.ts - Conversation agent → Costing and conversation log: logs exchange
src/agent.ts - Costing and conversation log → Runtime configuration: reads pricing
src/logger.ts - Terminal chat → Runtime configuration: reads settings
src/cli.ts - Conversation agent → Runtime configuration: reads model settings
src/agent.ts - HTTP server → Runtime configuration: reads server settings
src/server.ts - Health client → HTTP server: checks health
README.md