Charla docs
Charla is a platform to create, run and operate AI agents that people and organizations delegate work to through conversation: WhatsApp, Telegram and API.
Every page is Markdown. For coding agents: llms.txt, setup.md, openapi.json.
- Public API reference — Authentication, scopes, errors, idempotency and every developer endpoint
- Glossary — Shared vocabulary: agents, turns, lines, handoffs, WhatsApp terms
- Product vision — What Charla is, for whom, and its principles
- Product scope — Use-case catalog: personal, team, customer-facing and embedded agents
- Architecture — Contexts, containers, request flows and deployment topology
- Domain model — Entities, relationships, tenancy and IDs
- Agent runtime — The turn loop: inbound, steering, model, tools, delivery and memory
- Channels — WhatsApp, Telegram and the other channel adapters
- Tools, identity and governance — Webhook and MCP tools, identity, permissions, approvals and audit
- Evals and observability — Traces, eval cases, replay and signals
- Security, privacy and compliance — Encryption, tenant isolation, secrets and data protection
- Billing and plans — Credits, metering and plans