Chat & UI
The chat interface, rich rendering, navigation, and Temporary Chat.
Agent Tools
Image generation, weather, coding tools, prompt templates, and skills.
Models & Reasoning
Multi-model switching, the Anthropic (Claude), OpenAI (direct) and Microsoft Foundry providers, and the per-agent generation options (reasoning effort, with verbosity and the output budget derived from it).
Voice & Speech
Speech-to-text (REST and Realtime), text-to-speech providers, and Live full-duplex voice conversation with GPT-Live.
Knowledge & MCP
RAG over your PDFs, the Pipeline Jobs engine, and Model Context Protocol integration.
Memory & Sessions
Agent identity, a self-maintaining user preference memory, and session management with folders and auto titles.
Slash Commands
Run commands from the chat input with /help, /prompt, /skill, /model, completion, and dynamic arguments.
Cron Scheduler
Schedule scripts to run on a cron expression, an interval, or once after a delay -- managed from a portal, the API, or the agent.
File Explorer
A VSCode-style file tree and code editor for browsing and hand-editing files in your coding workspace.
Microsoft Teams
Talk to the ChatWalaʻau agent directly from Microsoft Teams -- personal chat, group chat, or a channel.
Declarative Agents
Define and switch the active agent with a YAML file or a GUI -- persona, model, per-agent tools, and output policy as data, no code change.
Declarative Workflows
Orchestrate multiple agents with a YAML workflow -- variables, conditionals, loops, agent, tool, HTTP, and human-in-the-loop actions compiled from a file, run in chat or as a background job.
Harness Agents
An autonomous software-engineering agent kind -- todo list, plan/execute modes, jailed file access, shell, skills, and web search -- composed in YAML and run as a chat run-target.
Webhook Gateway
Drive ChatWalaʻau from external events. The first source is Microsoft Graph, which auto-summarizes Teams meeting transcripts.
Ontology
Design concept models as RDF knowledge graphs on a visual canvas, search them with SPARQL or natural language, and let the assistant answer from them in any chat.
Token Usage Dashboard
See what your deployment consumed by day, month, type, model and chat, in your own time zone, and export the raw records for BI.