- Support
- Integrations
- Agent
Agent
AI integration · 17 node(s).
00Overview
Every field below is exactly what you see in the Flomation editor. Fields marked ● live picker let you choose from a list pulled live from your account — no IDs to look up.
01Connecting Agent
Agent nodes have no credential fields — there is no API key or secret to add as an environment secret. Fill in each node's own fields, listed below, and run the flow.
02Add
Add to Conversation
agent/add_to_conversation · Action
Store a message and return conversation history
| Field | Type | Details | |
|---|---|---|---|
| Conversation ID | string | Required | ${conversation_id} |
| Agent ID | string | Required | ${flow.agent_id} |
| Message Content | text | Required | |
| Role | string | Required | choices: User (inbound), Assistant (outbound) |
| Channel Type | string | ${channel_type} | |
| Sender | string | ${from} | |
| History Limit | integer | 30 |
Returns: tool_result, conversation_history, raw_history, message_id, success, error
03Create
Create Schedule
agent/create_schedule · Action
Create a recurring scheduled task for the agent
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | ${flow.agent_user_id} | |
| Conversation ID | string | ${flow.conversation_id} | |
| Schedule name | string | Required | |
| What to do when the schedule fires | string | Required | |
| Mode | string | Required | choices: Daily, Weekly, Interval |
| Time of day (HH:MM, for daily/weekly) | string | ||
| Days of week (comma-separated, for weekly) | string | ||
| Interval value (for interval mode) | string | ||
| Interval unit | string | choices: Minutes, Hours, Days | |
| Timezone (IANA) | string | Europe/London | |
| Source channel type | string | ${flow.channel_type} |
Returns: tool_result, schedule_id
04Delete
Delete Schedule
agent/delete_schedule · Action
Delete a scheduled task by name
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Schedule name to delete | string | Required |
Returns: tool_result
05Forget
Forget Memory
agent/forget · Action
Delete a specific memory from an agent's store
| Field | Type | Details | |
|---|---|---|---|
| Memory ID | string | Required | UUID of the memory to delete |
Returns: success
06Fulfill
Fulfil Commitment
agent/fulfill_commitment · Action
Mark an agent commitment as fulfilled after the follow-up has been delivered
| Field | Type | Details | |
|---|---|---|---|
| Commitment ID | string | Required | ${trigger.commitment_id} |
Returns: success
07Get
Get Calendar
agent/get_calendar · Action
Return the user's upcoming Google Calendar events. Use when the conversation hints at scheduling — where they are, where they need to be, whether they're running late.
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | Required | ${flow.agent_user_id} |
| How many hours ahead to look (default 24, max 168). | integer | 24 |
Returns: tool_result, events, event_count, no_calendar, success, error
Get Conversation
agent/get_conversation · Action
Fetch the full message history of a previous conversation the agent has had with this user. Use when a summary in your Recent Conversations section looks relevant.
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | Required | ${flow.agent_user_id} |
| Conversation ID — pass the conversation_id verbatim from the Recent Conversations section. | string | Required | 00000000-0000-0000-0000-000000000000 |
| Maximum messages to return (defaults to 200; max 500). Messages are read in sequence order from the start of the conversation. | integer | 200 |
Returns: tool_result, messages, message_count, returned_count, ended_at, was_truncated, success, error
08List
List Identities
agent/list_identities · Action
List all known channel identities for the current user
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | Required | ${flow.agent_user_id} |
Returns: tool_result, identities, count
List Schedules
agent/list_schedules · Action
List all scheduled tasks for the agent
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
Returns: tool_result, schedules, count
09Process
Process Extraction Output
agent/process_extraction · Action
Parse a structured extraction JSON payload and write memories, pending actions, and commitments
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Extraction JSON | string | Required | ${node.ai.response} |
| Agent User ID | string | ${trigger.agent_user_id} | |
| Conversation ID | string | ${trigger.conversation_id} | |
| Source Message ID | string | ${trigger.message_id} | |
| Channel Type | string | ${trigger.channel_type} |
Returns: memories_written, memories_flagged, memories_discarded, pending_actions_written, commitments_written, confirmations_processed, errors
10Read
Read Agent State
agent/read_state · Action
Read a persistent state value from an agent's key-value store
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ID of the agent |
| State Key | string | Required | conversation_history |
Returns: value, exists
11Recall
Recall Memories
agent/recall · Action
Fetch an agent's memories about a specific user
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | Required | ${flow.agent_user_id} |
| Pinned only | boolean | false | |
| Limit | integer | 20 | |
| Semantic search query | string | ||
| AWS region for embeddings | string | us-east-1 |
Returns: memories, count
12Remember
Remember Fact
agent/remember · Action
Store a durable fact, preference, or piece of feedback in an agent's memory
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | ${flow.agent_user_id} | |
| Scope | string | Required | choices: User-specific, Agent-global |
| Memory Type | string | Required | choices: Preference (auto-pinned), Feedback (auto-pinned), Fact, Relationship, Task, Session summary |
| Title | string | Required | Short handle for the memory |
| Body | string | Required | The fact itself |
| Pinned (always included in system prompt) | boolean | false | |
| Confidence (0.0–1.0, defaults 1.0) | string | 1.0 | |
| AWS region for embeddings | string | us-east-1 |
Returns: memory_id
13Search
Search Conversation History
agent/search_conversation · Action
Full-text search your entire message history with this user (all channels) to recall something outside your current context. Provide a search query.
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | Required | ${flow.agent_user_id} |
| Search query — words or a phrase to find in past messages. | string | Required | the demo wallpaper image |
| Max results | integer | 20 |
Returns: tool_result, results, match_count, success, error
14Send
Send Agent Message
agent/send_message · Action
Send a message through an agent's communication channel and record it
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ID of the agent |
| Channel | string | Required | choices: Telegram, Email, Webhook |
| Recipient | string | Required | Chat ID, email address, or webhook URL |
| Message | text | Required | Message content |
Returns: message_id, success
15Set
Set Reminder
agent/set_reminder · Action
Schedule a follow-up. The platform resolves the timing and returns the exact date it stored.
| Field | Type | Details | |
|---|---|---|---|
| When to fire, in plain English: a duration ("in 72 hours", "in 30 minutes"), a relative day ("tomorrow at 9am", "next Monday"), or a date ("1 October", "25 December at 09:00"). Never work the calendar date out yourself — pass what was asked for and use the date this returns. | string | Required | in 72 hours |
| What to remind about, written so it still makes sense on its own days later | string | Required | Pull the updated campaign numbers and compare the click-to-view ratio |
| What sort of follow-up this is | string | choices: Reminder, Follow-up, Monitor, Chase | |
| Leave empty for a one-off. Set it only when the follow-up was asked for repeatedly | string | choices: One-off, Daily, Weekdays, Weekly, Monthly | |
| Agent ID | string | Required | ${flow.agent_id} |
| Agent User ID | string | ${flow.agent_user_id} | |
| Conversation ID | string | ${flow.conversation_id} |
Returns: tool_result, success, commitment_id, due_at, due_at_friendly
16Write
Write Agent State
agent/write_state · Action
Write a persistent state value to an agent's key-value store
| Field | Type | Details | |
|---|---|---|---|
| Agent ID | string | Required | ID of the agent |
| State Key | string | Required | conversation_history |
| Value | object | Required |
Returns: success
17Notes & Limitations
Behaviours and constraints worth knowing before you build with these nodes.
- Memories are scoped to a single user, so a fact stored for one person is not returned when the agent recalls memories while serving a different user.
- Scheduled tasks are addressed by their name, so the name you set when creating one is the handle you must supply later to delete it.
- Adding a message to a conversation also returns the full running history, so the same step doubles as a read of the thread.
- Retrieving calendar events requires the user's Google Calendar to be connected first and returns only upcoming events, not past ones.