Skip to main content

15 docs tagged with "functions"

View all tags

Agent function examples

Complete agent functions you can adapt. Each table, column or API drawn as a chip, such as orders, orders.email or createTicket, is a reference you insert from the editor's suggestions. When you rebuild an example, pick your own tables, columns and APIs: type " after tables.search( or apis.execute(, or "@ anywhere else. See Get suggestions.

Agent functions

An agent function is a short piece of JavaScript that your AI agent runs when it needs custom logic: looking up a record, calling one of your systems, or preparing data before the agent replies. From its code, an agent function can work with your tables, call the APIs configured for your agent, read and update global variables, and, on runtime 2026.10.19, send messages to the user while it works.

Call APIs from agent functions

This page covers calling your APIs from the code of an agent function with apis.execute from "@yellow/bot". The APIs themselves are set up under Build → API. Credentials such as API keys stay in the API's settings and never appear in your code.

Crypto helpers in agent functions

crypto from "@yellow/bot" creates IDs, hashes and signs values, and encrypts and decrypts with AES. These helpers don't count towards any limit, and they return their result directly, so no await is needed.

Halt the agent's turn

Check or change a function's version in the editor's status bar. See Runtime versions.

Handle errors in agent functions

A function fails when its code throws, when a call it makes fails and isn't caught, or when it goes past a limit. This page covers what the agent or flow sees then, the errors you're likely to meet, and how to keep the conversation going.

JavaScript in agent functions

Agent functions run modern JavaScript in a secure sandbox, not in Node.js or a browser. This page lists what the language gives you, the built-ins added for you, and what's blocked.

Runtime versions

Every agent function runs on a dated runtime version: the set of features its code can use, such as which modules it can import. A function's version never changes on its own. You move it to a newer version when you want what that version adds, and you can move it back.

Test and debug agent functions

Test an agent function with the Run button while you build it, and add console logs to see what it did, both in the editor and in live conversations.

Use global variables in agent functions

Global variables hold information the rest of the conversation can use, such as the customer's tier, their language, or the last order they asked about. An agent function can read them to make decisions, and update them so that later steps (other agents, flows or functions) can use what it found.

Work with tables in agent functions

This page covers using your agent's database tables from the code of an agent function: searching, inserting and updating rows with tables from "@yellow/bot". You create the tables and their columns on the Database page.

Work with voice calls

Check or change a function's version in the editor's status bar. See Runtime versions.