polite.aiDocs
BUILD / Agents & teams

The agent editor

Every agent opens in a drawer — tabs covering its identity, voice, conversation, tools and deployments, all autosaving as you type.

Tap any agent — a card or member pill on the Agents page, or a node on the Builder canvas — and its editor slides over the current screen. No navigation, no losing your place; drag the edge to resize it and the width is remembered.

Autosave: there is no Save button

Every change autosaves into the team's draft — the footer ticks through autosaving… to autosaved 12 s ago so you can watch it happen. Real callers stay on the live version until you go live. Just opening the drawer never creates a draft; only actual edits do, so browsing is always safe.

Overview, model and voice

Overview holds the agent's name, description, type and team membership, plus a Test in browser button and Download config for the full configuration as JSON.

Model & voice (or Model & behaviour for text agents) is where you pick the AI model from a searchable picker and set the temperature slider between Precise and Creative — the detail is in Models & platforms. For voice agents the same tab cascades through Language, Voice (each shown as name · vendor) and Speech recognition; changing the language reloads the voices and recognisers that support it.

Conversation: greeting and prompt

The Conversation tab controls how calls open — No greeting, Say this exactly for verbatim text, or Generate from instructions to let the model write it — and holds the agent's full System prompt in a Markdown editor. This is the hand-edit surface for wording you care about precisely.

Tools & transfers

Three sub-tabs for the power user: Functions (custom tools, including REST calls with typed parameters), API keys (named secrets, write-only once saved — they can never be read back), and MCP servers for connecting external tool servers. Tools wired by the team graph — transfers and subagents — appear read-only, labelled Managed by the team graph — edit in the builder: change the team's structure in the Builder chat, not here.

Deploy and Activity

Deploy lists everywhere the agent is answering, each with an Undeploy link, and a Deploy to form targeting the browser, an owned number or a SIP registration. It's also home to the telephone test and a Place an outbound call section — see Outbound calling.

Activity shows the agent's most recent calls at a glance; the full, filterable history lives on the Calls page.

Hand-edit or ask the Builder?

Both write to the same draft, so it's purely a question of ergonomics. Use the drawer when you know exactly what you want: this greeting text, that voice, a one-line prompt change. Ask the Builder chat when the change is structural — new agents, rewiring transfers, new tools — or when you'd rather describe the outcome and let it work out the edits.

Good to know

Delete lives in the drawer footer behind a confirm — and deleting is the one thing autosave can't undo, so it asks properly first.

Last updated 2026-07-30