Docs for building agentic
conversation at scale.
Every guide, API endpoint, and playbook for the Yellow.ai Agentic Platform — one search box away.
Where to start
Meet the AI layer
Build and run your Nexus bot by asking — the in-app assistant does the work.
See what it can doBuild a v3 agent
Super agent, agents, and tools — how the pieces fit and how to configure them.
Start buildingUse the API
Auth, your first call, webhooks, events, and the OpenAPI spec.
Read the referenceThe rest of the docs
Recently shipped
Full changelogTesting Lab, rebuilt around traces
A redesigned Testing Lab that shows execution traces right beside the conversation, with AI-generated test scenarios and one-click test creation from real chats.
AI Trust Centre meets Nexus
Ask Nexus for your coverage report, fix failing test cases with a click, and watch overview metrics fill in — the Trust Centre is now conversational.
Chat widget v3 upgrades
The v3 chat widget picks up themes and primary colors, right-to-left support for Arabic, configurable scroll behaviour, and a setting to open in voice or chat mode by default.
Voice goes GA-grade
Realtime voice moved to the GA API with a new default voice, clearer text-to-speech delivery, and test calls you can place straight from the platform.
Overview drill-downs and conversation logs polish
Click into contained and non-contained conversations from the overview, follow a resolution status trend, and review logs in a non-modal drawer that keeps your place.
Stuck?
The community is the fastest place to get unstuck. Otherwise, [email protected] picks things up within a business day.