BotStar

Manage BotStar chatbots from a flow, publishing drafts, editing bot and user attributes, and maintaining CMS content collections. Agents keep conversational content current and send Facebook Messenger messages on cue.

25 actions API key available

Platform Documentation ↗ Capability data verified 2026-08-02

A product feed or FAQ source changes upstream

Agent reads the collection schema with Get CMS Entity to match incoming fields

Agent lists current items with List CMS Entity Items to decide what is new and what is a change

Agent writes every item Draft Only with Create CMS Entity Item and Update CMS Entity Item

Agent re-reads the draft environment to confirm each row landed as intended

Content owner receives the staged diff plus anything else already sitting in the draft

Owner approves before Publish Bot releases the draft to live

What This Integration Enables

BotStar is the outlier in the chatbot category, and the reason is architectural. Most bot platforms store their answers as configuration inside the flow builder, which means changing an answer means editing a bot. BotStar stores them as content: CMS entities are structured collections with their own field schemas and items, so a product catalogue or an FAQ list lives as records rather than as message blocks. On top of that sits a genuine staging model. Reads choose Draft or Live. Writes choose Draft Only or Draft And Live. Publish Bot is the release. If you have ever wanted a bot whose answers can be maintained by the same pipeline that maintains your website copy, this is the platform built for it.

FlowRunner agents use both halves. On the content side they list, create, read, rename, and delete CMS entities, add and update and remove field definitions in batch, and page through items to create, read, update, and delete individual records with values keyed by each field's unique name. On the conversation side they list and create bots and read one back to confirm an ID still resolves. They manage bot-level attributes, which hold values shared across every conversation such as business hours or a support address. They extend the end-user profile schema with custom attributes, read and partially update an end user's profile, and send a Facebook Messenger message through the bot the token belongs to. Every write that touches content or schema carries the environment choice with it, so an agent decides explicitly whether a change is staged or live.

Without FlowRunner

Bot answers edited one block at a time Somebody opens the editor and retypes a price into every place it appears

No separation between editing and releasing A change made to fix one answer goes live the moment it is saved

Catalogue and bot drift apart The product list the bot quotes is a copy that stopped matching the source months ago

With FlowRunner

Content maintained as structured records Entities, fields, and items are written from the source system by unique name

Staging is the default write path Draft Only holds every change until a release is deliberately made

Release is one reviewed decision Publish Bot runs after a person has read what the draft contains

Use Case Scenarios

Catalogue Sync From the Source of Record

The product list lives in Airtable and changes weekly. On a schedule the agent reads the BotStar entity schema with Get CMS Entity, pages through the existing records with List CMS Entity Items, and works out the delta. New products become Create CMS Entity Item calls, changed ones become partial updates by unique name, and discontinued ones get flipped to disabled with Update CMS Entity Item rather than deleted. Everything is written Draft Only. The bot keeps quoting last week's catalogue until somebody releases the new one, which is the correct behavior for a bot that quotes prices.

AI Rewrite of Product Descriptions, Reviewed Before Release

The team wants shorter, more consistent product blurbs. The agent pulls the current items with List CMS Entity Items, passes each description through an AI step with the house style guide as context, and writes the results back with Update CMS Entity Item against the Draft environment only. Nothing the model wrote is visible to a customer yet. The content owner reads the staged set, edits the handful that missed, and releases the batch. Machine-written copy reaches customers only after a person has read it.

Shared Values Updated Without Touching Flows

Support hours change, and the bot mentions them in nine different blocks. Because those blocks read a bot attribute rather than a hardcoded string, the agent calls Update Bot Attribute once with the new value. The partial update preserves the description, so a flow that only knows the new hours does not have to supply anything else. One write changes every place the bot says it.

Human-in-Loop Highlight

Publish Bot is the operation to be careful with, and not for the reason people expect. The risk is not that the agent's own changes are wrong. It is that publishing is a release of the entire draft, so anything a colleague staged in the BotStar editor since the last publish ships alongside whatever the agent staged, to every user on every connected channel, in one call. An automated catalogue sync that publishes on its own is quietly deciding to release somebody else's half-finished work. FlowRunner puts a human-in-the-loop gate exactly there. The agent writes Draft Only, re-reads the draft environment to confirm its own writes, then stops and asks the content owner through Slack: "Catalogue sync staged in Draft: 42 items updated, 6 created, 3 disabled. The draft also contains 2 attribute changes and 1 new FAQ item this flow did not make. Publish the whole draft to live, or hold?" The owner publishes, or goes and looks first. The other reason to keep a person here is the neighbouring delete: Delete CMS Entity removes the collection with its field definitions and every item in it, and BotStar cannot undo that through the API, so an agent proposing one shows the item count and waits.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via Slack

Human decides

Agent resumes with decision

Agent Capabilities

25 actions

Bots

List Bots Returns every chatbot the API token can access with its ID, name, and owning team. The bot ID is required by almost every other operation, so this is normally the first call.
Create Bot Creates a new, empty chatbot in the account and returns its generated ID. The conversation flow is built afterwards in the BotStar editor and released with Publish Bot.
Get Bot Retrieves a single bot by ID with its name and owning team. Used to confirm a stored bot ID still resolves before running a longer sequence against it.
Publish Bot Publishes the bot's draft version to live, making every pending change visible to end users on the connected channels. Run after a batch of draft-only content or attribute updates to release them together.

Bot Attributes

List Bot Attributes Returns the bot-level attributes defined for the selected environment. These hold values shared across every conversation, such as business hours or a support email, and are read inside the bot's blocks.
Create Bot Attribute Creates a bot-level attribute in the chosen environment. Draft Only stages it for a later publish; Draft And Live makes it available to running conversations immediately.
Update Bot Attribute Applies a partial update to an existing bot attribute, requiring at least one of description or value. Fields left empty are preserved, so a flow that only knows the new value can call it safely.
Delete Bot Attribute Deletes a bot attribute from the chosen environment. Draft Only leaves the live bot untouched until the next publish; Draft And Live removes it from running conversations straight away.

End Users

Get Bot User Retrieves the full profile BotStar holds for one end user, including name, email, birthday, channel, locale, timezone offset, profile picture, IP address, tags, and any one-time notification topics they subscribed to.
Update Bot User Applies a partial update to an end user's profile, requiring at least one field. Custom attributes defined on the profile schema can be set in the same call.
Create User Attribute Defines a new custom attribute on the bot's end-user profile schema so every user can carry that field. Once created it can be written per user and referenced inside the bot's blocks.

CMS Entities

List CMS Entities Returns the CMS entities defined on a bot for the selected environment. Entities are BotStar's structured content collections, such as a product catalogue or an FAQ list, each owning a set of fields and items.
Create CMS Entity Creates a CMS entity together with its initial field definitions. Each field needs a unique name, a display name, and a data type, with option-based fields carrying their selectable values.
Get CMS Entity Retrieves one CMS entity by ID with its full field schema. Read it before writing items so a flow can match incoming data onto the entity's unique names and data types.
Update CMS Entity Renames an existing CMS entity. Fields and items are untouched, so a collection can be relabelled without disturbing the content stored in it.
Delete CMS Entity Deletes a CMS entity together with its field definitions and every item stored in it. This cannot be undone through the API, so the entity ID is confirmed before the call is automated.

CMS Fields

Create CMS Entity Fields Adds one or more new fields to an existing entity in a single call, each with a unique name, display name, and data type, plus choices for option-based fields.
Update CMS Entity Fields Updates one or more existing fields in a single call, matched by unique name, so a field label can be renamed or an option list extended without recreating the entity.
Delete CMS Entity Fields Removes one or more fields from an entity by unique name. Draft Only stages the schema change until the next publish; Draft And Live applies it to the running bot immediately.

CMS Items

List CMS Entity Items Returns the content items stored in a CMS entity. This is the only paginated BotStar resource, defaulting to 100 per page, and can be narrowed by item name and by enabled or disabled status.
Create CMS Entity Item Creates a content item inside an entity with its name, status, and field values keyed by each field's unique name. Multiple-value and multiple-option fields accept an array of strings.
Get CMS Entity Item Retrieves a single item by ID from the chosen environment with its name, status, and every stored field value. Used to read one catalogue or FAQ record before deciding what a flow should do next.
Update CMS Entity Item Applies a partial update to an item: change its name, flip it between enabled and disabled, or overwrite individual field values by unique name. Anything omitted is left as it was.
Delete CMS Entity Item Deletes a single item from an entity. Draft Only stages the removal until the next publish; Draft And Live takes the item out of the running bot straight away.

Messaging

Send Facebook Message Sends a message to a Facebook Messenger user through the bot the API token belongs to, so no bot ID is needed. The body uses BotStar's own message format, and a message tag states the reason when the send falls outside Facebook's 24-hour messaging window.

Frequently Asked Questions

What can FlowRunner do with BotStar?

FlowRunner agents can run List Bots, Create Bot, and Get Bot in BotStar, plus 22 more actions.

Does connecting BotStar to FlowRunner require OAuth?

No. BotStar connects to FlowRunner with an API key, no OAuth flow required.

Can BotStar trigger a FlowRunner workflow automatically?

BotStar doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with BotStar

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us