Relevance AI
App · 17 actions
Relevance AI agent and tool execution, tasks and knowledge.
Actions
| Action | What it does |
|---|---|
| Delete agent | Delete agent. |
| Delete knowledge rows | Delete knowledge rows. |
| Delete knowledge set | Delete knowledge set. |
| Get agent | Get agent. |
| Get tool | Get tool. |
| Get tool job | Get tool job. |
| List agent tools | List agent tools. |
| List agents | List agents. |
| List knowledge rows | List knowledge rows. |
| List knowledge sets | List knowledge sets. |
| List task tool runs | List task tool runs. |
| List tasks | List tasks. |
| List tools | List tools. |
| Run tool | Run tool. |
| Start tool job | Start tool job. |
| Trigger agent | Trigger agent. |
| Upsert agent | Upsert agent. |
Delete agent
Delete agent.
Inputs
- Agent
agent_idtextRequired - Agent id.
Outputs
dataobject- Api result.
Example output
{
"data": {}
}Delete knowledge rows
Delete knowledge rows.
Inputs
- Knowledge set
knowledge_settextRequired - Knowledge set.
- Document id
document_idtext - Document id.
- Filters
filterslist - Currently limited to exact_match, ids, exists, regexp filter_type
Outputs
dataobject- Api result.
Example output
{
"data": {}
}Delete knowledge set
Delete knowledge set.
Inputs
- Knowledge set
knowledge_settextRequired - Knowledge set.
Outputs
dataobject- Api result.
Example output
{
"data": {}
}Get agent
Get agent.
Inputs
- Agent
agent_idtextRequired - Agent id.
- Version
versiontext - Version of the agent to run. Can be 'active', 'draft', or a specific version_id
Outputs
dataobject- Api result.
Example output
{
"data": {
"agent": {
"_id": "example",
"accept_attachments_from_parent": false,
"action_behaviour": "example",
"action_retry_config": {
"after_retries_behaviour": "terminate-conversation",
"force_retry": false,
"max_retries": 1.0
},
"actions": [],
"active_version_id": "example",
"agent_card": {
"flavor_text": "example",
"generated_at": 1.0,
"image_url": "example"
},
"agent_decide_prompt": "example",
"agent_id": "example",
"always_create_new_conversation_when_subagent": false,
"autonomy_limit": 1.0,
"autonomy_limit_behaviour": "ask-for-approval",
"capabilities": {
"computer": {
"enabled": false,
"github": {
"allow_push_from_triggers": false,
"enabled": false,
"repositories": []
},
"internet_access": "none",
"size": "standard"
},
"web_fetch": {
"enabled": false
},
"web_search": {
"enabled": false
}
},
"categories": [],
"conditional_approval_rules": {
"max_approvals_asked": 1.0,
"max_auto_approvals": 1.0
},
"conversation_compression_enabled": false,
"creation_surface": "inventor",
"custom_metadata": {
"options": [],
"read_metadata_enabled": false,
"read_metadata_system_prompt": "example",
"system_prompt": "example"
},
"default_eval_config": {
"block_on_failure": false,
"test_set_ids": [],
"threshold_score": 1.0
},
"deleted_reason": "entity_deleted",
"dependency_config": {
"mode": "legacy"
},
"description": "example",
"draft_version_id": "example",
"embed_config": {},
"embeddable": false,
"emoji": "example",
"enable_python_executor": false,
"escalations": {
"email": {
"emails": "example"
},
"slack": {
"channels": []
}
},
"expiry_date_": "example",
"external_message_relay_config": [],
"fallbackModel": "openai-gpt-4o",
"in_marketplace": false,
"insert_date_": "example",
"is_pinned": false,
"is_scheduled_triggers_enabled": false,
"knowledge": [],
"last_updated_by": {
"user_id": "example",
"user_name": "example"
},
"machine_user_id": "example",
"mas_id": "example",
"max_job_duration": "hours",
"memory": {
"add_memory_prompt": "example",
"compaction": {
"mode": "off",
"tool_output": "pass_full"
},
"delete_memory_prompt": "example",
"enable_delete_tool": false,
"enabled": false,
"memory_level": "project"
},
"metadata": {
"clone_count": 1.0,
"custom_task_views": [],
"last_run_date": "example"
},
"mode": {
"external_relay": {
"config": {
"channel_id": "example",
"oauth_account_id": "example"
},
"type": "example"
},
"meeting_bot_config": {
"meeting_bot_image": "example",
"recording_mode": "Audio Only",
"timeout_settings": {
"everyone_left_timeout": 1.0,
"max_recording_time": 1.0,
"noone_joined_timeout": 1.0,
"waiting_room_timeout": 1.0
},
"transcription_model": "Default"
},
"summary_prompt": "example",
"type": "example",
"version": 1
},
"model": "openai-gpt-4o",
"model_options": {
"max_output_tokens": 1.0,
"parallel_tool_calls": false,
"strict_mode": false,
"switch_model_after_n_tokens": {
"model": "example",
"n_tokens": 1.0
}
},
"name": "example",
"notifications_config": {
"notifications": []
},
"origin": {
"marketplace_listing": {
"author_org_display_id": "example",
"author_project_display_id": "example",
"author_user_display_id": "example",
"can_clone": false,
"can_republish": false,
"display_id": "example",
"entity_display_id": "example",
"entity_project_display_id": "example",
"entity_region": "example",
"entity_type": "agent",
"entity_version_display_id": "example"
},
"project": "example",
"region": "1e3042",
"root_agent_id": "example"
},
"params": {},
"params_schema": {
"metadata": {
"field_order": []
},
"properties": {}
},
"project": "example",
"prompt_description": "example",
"public": false,
"remote_mcp_configs": {},
"runner": {
"type": "example"
},
"runtime": {
"code": "example",
"enabled": false,
"engine": "default",
"meeting": {
"background_denoising_enabled": false,
"first_message": "example",
"first_message_interruptions_enabled": false,
"first_message_mode": "agent-speaks-first",
"idle_messages": [],
"idle_timeout_seconds": 1.0,
"max_duration_seconds": 1.0,
"silence_timeout_seconds": 1.0,
"summary_prompt": "example",
"transcriber": {
"language": "example",
"model": "example",
"provider": "example"
},
"voice": {
"language": "example",
"model": "example",
"provider": "example",
"speed": 1.0,
"voiceId": "example"
}
},
"phone_call": {
"background_denoising_enabled": false,
"background_sound": "off",
"end_call_tool_enabled": false,
"engine_version": "v1",
"first_message": "example",
"first_message_interruptions_enabled": false,
"first_message_mode": "agent-speaks-first",
"idle_messages": [],
"idle_timeout_seconds": 1.0,
"inbound_phone_number": {
"numberId": "example",
"type": "example"
},
"keypad_input": {
"delimiter": "#",
"enabled": false,
"timeout_seconds": 1.0
},
"max_duration_seconds": 1.0,
"outbound_phone_number": {
"numberAlias": "US1",
"type": "example"
},
"recording_enabled": false,
"silence_timeout_seconds": 1.0,
"summary_prompt": "example",
"transcriber": {
"language": "example",
"model": "example",
"provider": "example"
},
"transfer_call_tool": {
"destinations": [],
"enabled": false
},
"voice": {
"language": "example",
"model": "example",
"provider": "example",
"speed": 1.0,
"voiceId": "example"
},
"voicemail_message": "example"
},
"preset": {
"preset_agent_id": "inventor"
},
"webhook": "example"
},
"share_styles": {
"bubble_icon": "example",
"bubble_style": "agent",
"hide_conversation_list": false,
"hide_description": false,
"hide_file_uploads": false,
"hide_logo": false,
"hide_tool_steps": false,
"input_placeholder_text": "example",
"primary_color": "example",
"selected_format": "shareable-link",
"starting_message_prompts": []
},
"starting_messages": [],
"suggest_replies": false,
"suggest_replies_prompt": "example",
"suggested_prompts": [],
"suggested_triggers": [],
"sync_type_settings": [],
"system_prompt": "example",
"tags": {},
"temperature": 1.0,
"template": {
"agent_id": "example",
"project": "example",
"region": "1e3042"
},
"thinking": {
"_oneof_type_": "example",
"reasoning_effort": "none"
},
"thinking_tool": {
"enabled": false
},
"title_prompt": "example",
"triggers": [],
"type": "default",
"update_date_": "example",
"use_streaming": false,
"user_instructions": "example",
"user_tag_ids": [],
"verbosity": {
"_oneof_type_": "example",
"verbosity": "low"
},
"version": "example",
"views": 1.0,
"visual_data_masking": {
"enabled": false
}
}
}
}Get tool
Get tool.
Inputs
- Studio
studio_idtextRequired - Studio id.
- Version
versiontext - Version.
- Tool version
tool_versiontext - The version of the tool, can be 'active', 'draft' or a specific version_id
- Max job duration
max_job_durationtext - Defines job execution strategy. 'hours' routes to a dedicated worker tier with a concurrency of 1 and slower autoscaling , only use when jobs genuinely need >1h runtime. 'minutes' (default) uses the standard tier with higher concurrency.One of:
hours,minutes,synchronous_seconds(Synchronous seconds),background_seconds(Background seconds). - Job id
job_idtext - Can be used to control the job id.
Outputs
dataobject- Api result.
Example output
{
"data": {
"studio": {
"_id": "example",
"active_version_id": "example",
"contains_legacy_transformation": false,
"cover_image": "example",
"creation_surface": "inventor",
"creator_display_picture": "example",
"creator_first_name": "example",
"creator_last_name": "example",
"creator_user_id": "example",
"deleted_reason": "entity_deleted",
"description": "example",
"display_output": {
"content": "example",
"is_safe": false,
"type": "html"
},
"draft_version_id": "example",
"emoji": "example",
"in_marketplace": false,
"insert_date_": "example",
"integration_requirements": {
"api_keys": [],
"oauth": []
},
"is_hidden": false,
"machine_user_id": "example",
"max_job_duration": "hours",
"metadata": {
"clone_count": 1.0,
"clone_count_percentage": 1.0,
"last_run_date": "example",
"legacy_mcp_action_ids": [],
"source_marketplace_listing": {
"author_org_display_id": "example",
"author_project_display_id": "example",
"author_user_display_id": "example",
"can_clone": false,
"can_republish": false,
"display_id": "example",
"entity_display_id": "example",
"entity_project_display_id": "example",
"entity_region": "example",
"entity_type": "agent",
"entity_version_display_id": "example"
},
"source_project": "example",
"source_region": "example",
"source_studio_id": "example"
},
"metrics": {
"executions": 1.0,
"views": 1.0
},
"output_schema": {
"metadata": {
"field_order": []
},
"properties": {}
},
"owner_user_id": "example",
"params_schema": {
"metadata": {
"field_order": []
},
"properties": {}
},
"predicted_output": [],
"project": "example",
"prompt_description": "example",
"public": false,
"publicly_triggerable": false,
"region": "example",
"schedule": {
"frequency": "hourly"
},
"share_id": "example",
"share_styles": {
"cta_icon": "example",
"cta_text": "example",
"guidance_tip_text": "example",
"hide_guidance_tip": false,
"hide_logo": false,
"primary_color": "example",
"selected_format": "shareable-link"
},
"state": {
"params": {},
"steps": {}
},
"state_mapping": {},
"studio_id": "example",
"tags": {
"categories": {},
"integration_source": "example",
"integrations": {},
"type": "transformation",
"use_cases": {}
},
"title": "example",
"transformations": {
"output": {},
"steps": []
},
"update_date_": "example",
"version": "example"
}
}
}Get tool job
Get tool job.
Inputs
- Studio
studio_idtextRequired - Studio id.
- Job id
job_idtextRequired - Job id.
- After message id
after_message_idnumber - After message id.
- Filters
filterslist - Filters.
- Sort
sortlist - Fields to sort by. For each field, sort by descending or ascending. If you are using descending by datetime, it will get the most recent ones. Can be provided in format ["sortfield","subsortfield"] or in format [{"sortfield":"desc"}, {"subsortfield":"asc"}]. If provided as a query parameter, use: sort=${JSON.stringify([{"sortfield":"desc"}])}
- Page size
page_sizenumber - Page size.
- Ending update only
ending_update_onlytrue/false - Ending update only.
- Project
projecttext - Project.
Outputs
dataobject- Api result.
Example output
{
"data": {
"last_message_id": 1.0,
"type": "timeout",
"updates": []
}
}List agent tools
List agent tools.
Inputs
- Agent
agent_idtextRequired - Agent id.
- Agent override
agent_overrideobject - Agent override.
- Include phantom tools
include_phantom_toolslist - Include phantom tools.
- Agent version
agent_versiontext - Version of the agent to run. Can be 'active', 'draft', or a specific version_id
- Include unconnected tools
include_unconnected_toolstrue/false - Also list the tools of per-user connectors the caller has not connected, from the tool list saved on the agent, each with `missing_connection` set. They cannot run until the caller connects.
Outputs
dataobject- Api result.
Example output
{
"data": {
"chains": []
}
}List agents
List agents.
Inputs
- Query
querytext - Query.
- Filters
filterslist - Filters.
- Page
pagenumber - Page.
- Page size
page_sizenumber - Page size.
- Sort
sortlist - Fields to sort by. For each field, sort by descending or ascending. If you are using descending by datetime, it will get the most recent ones. Can be provided in format ["sortfield","subsortfield"] or in format [{"sortfield":"desc"}, {"subsortfield":"asc"}]. If provided as a query parameter, use: sort=${JSON.stringify([{"sortfield":"desc"}])}
- Include public agents
include_public_agentstrue/false - DEPRECATED: this parameter had no effect and is now ignored.
- Agent versions
agent_versionslist - Agent versions.
Outputs
dataobject- Api result.
Example output
{
"data": {
"results": []
}
}List knowledge rows
List knowledge rows.
Inputs
- Query
querytext - Query.
- Knowledge set
knowledge_settextRequired - Use * as knowledge set to retrieve all knowledge sets
- Document id
document_idtext - Document id.
- Filters
filterslist - ONLY APPLIES WHEN `query` ISN'T PROVIDED. Currently limited to exact_match, ids, exists, regexp filter_type
- Page
pagenumber - Page.
- Page size
page_sizenumber - Page size.From 1 to 1000.
- Include fields
include_fieldslist - Fields to include in the response. Default is all fields
- Sort
sortlist - Fields to sort by. For each field, sort by descending or ascending. If you are using descending by datetime, it will get the most recent ones. Can be provided in format ["sortfield","subsortfield"] or in format [{"sortfield":"desc"}, {"subsortfield":"asc"}]. If provided as a query parameter, use: sort=${JSON.stringify([{"sortfield":"desc"}])}
Outputs
dataobject- Api result.
Example output
{
"data": {
"results": []
}
}List knowledge sets
List knowledge sets.
Inputs
- Filters
filterslist - Currently limited to exact_match, ids, exists, regexp filter_type
- Sort
sortlist - Fields to sort by. For each field, sort by descending or ascending. If you are using descending by datetime, it will get the most recent ones. Can be provided in format ["sortfield","subsortfield"] or in format [{"sortfield":"desc"}, {"subsortfield":"asc"}]. If provided as a query parameter, use: sort=${JSON.stringify([{"sortfield":"desc"}])}
- Include metadata
include_metadatatrue/false - Include metadata.
- Include hidden
include_hiddentrue/false - Include hidden.
- Include counts
include_countstrue/false - Include counts.
- Page
pagenumber - Page.
- Page size
page_sizenumber - Page size.
- Query
querytext - Query.
Outputs
dataobject- Api result.
Example output
{
"data": {
"results": []
}
}List task tool runs
List task tool runs.
Inputs
- Agent
agent_idtextRequired - Agent id.
- Conversation id
conversation_idtextRequired - Conversation id.
- Page size
page_sizenumber - Page size.
- Page
pagenumber - Page.
Outputs
dataobject- Api result.
Example output
{
"data": {
"agentDetails": {},
"results": [],
"studios": {}
}
}List tasks
List tasks.
Inputs
- Page
pagenumber - Page.
- Page size
page_sizenumber - Page size.
- Include agent details
include_agent_detailstrue/false - Include agent details.
- Include eval details
include_eval_detailstrue/false - Include eval details.
- Include debug info
include_debug_infotrue/false - Include debug info.
- Exclude heavy params
exclude_heavy_paramstrue/false - Exclude heavy params.
- Query
querytext - Query.
- Mask pii
mask_piitrue/false - Mask pii.
- Filters
filterslist - Filters.
- Sort
sortlist - Fields to sort by. For each field, sort by descending or ascending. If you are using descending by datetime, it will get the most recent ones. Can be provided in format ["sortfield","subsortfield"] or in format [{"sortfield":"desc"}, {"subsortfield":"asc"}]. If provided as a query parameter, use: sort=${JSON.stringify([{"sortfield":"desc"}])}
- Event logs filters
event_logs_filterslist - Filter conversations by event type and trigger type/studio id
- Source
sourceobject - List the conversations this source started, newest first, paged by `cursor` instead of `page`. `alarm` and `inventor_routine` require loops-list-be, and `alarm` also needs project-scoped-eval-observability-be. Cannot be combined with `page`, `sort`, `query`, `filters` or `event_logs_filters`.
- Cursor
cursortext - With `source`: the previous page's `next_cursor`. Omit for the first page.
- Exclude fields
exclude_fieldslist - Optional list of fields to omit from each returned row (dot-paths within the returned metadata, e.g. `conversation.debug`, `conversation.pending_approvals`). Any field not listed is always returned; omit the param for the full response. Excluding expensive fields (the heavy `conversation.*` blobs , e.g. `conversation.debug`, `conversation.workforce_context`, `conversation.pending_approvals`) improves latency; excluding cheaper fields simply removes them from the response.
Outputs
dataobject- Api result.
Example output
{
"data": {
"agents": {},
"costs": {},
"evals": {},
"next_cursor": "example",
"results": []
}
}List tools
List tools.
Inputs
- Filters
filterslist - Filters.
- Select fields
select_fieldslist - Fields to include in the search results, empty array/list means all fields.
- Page size
page_sizenumber - Size of each page of results
- Page
pagenumber - Page of the results
- Sort
sortlist - Fields to sort by. For each field, sort by descending or ascending. If you are using descending by datetime, it will get the most recent ones. Can be provided in format ["sortfield","subsortfield"] or in format [{"sortfield":"desc"}, {"subsortfield":"asc"}]. If provided as a query parameter, use: sort=${JSON.stringify([{"sortfield":"desc"}])}
- Include vector
include_vectortrue/false - Include vectors in the search results.
- After id
after_idlist - Retrieve items after these sort values in the sort order.
- Query
querytext - Query.
- Version
versiontext - Version.
- Tool version
tool_versiontext - The version of the tool, can be 'active', 'draft' or a specific version_id
- Max job duration
max_job_durationtext - Defines job execution strategy. 'hours' routes to a dedicated worker tier with a concurrency of 1 and slower autoscaling , only use when jobs genuinely need >1h runtime. 'minutes' (default) uses the standard tier with higher concurrency.One of:
hours,minutes,synchronous_seconds(Synchronous seconds),background_seconds(Background seconds). - Job id
job_idtext - Can be used to control the job id.
- Return openapi schema
return_openapi_schematrue/false - Return openapi schema.
- Include metrics
include_metricstrue/false - Include metrics.
- Include share links
include_share_linkstrue/false - Include share links.
- Include version metadata
include_version_metadatatrue/false - Include version metadata for each transformation step (current version, latest version, deprecation info). Defaults to false to avoid performance overhead.
- Tool versions
tool_versionslist - Tool versions.
- Created by user
created_by_usertrue/false - When true, restrict results to studios where creator_user_id matches the authenticated user. Applied inside the DoubleListTool double-read path.
- Screens
screenslist - This attribute lets you choose which screens an item will appear on, filtering in. Defaults to all screens. Use hidden to remove from everywhere.One or more of:
notebook,workforce,agents.
Outputs
dataobject- Api result.
Example output
{
"data": {
"openapi_schema": {},
"results": []
}
}Run tool
Run tool.
Inputs
- Studio
studio_idtextRequired - Studio id.
- Project
projecttext - This can be provided to execute a studio in a project different to the one specified in the auth header.
- Params
paramsobjectRequired - Params.
- Version
versiontext - Version.
- Tool version
tool_versiontext - The version of the tool, can be 'active', 'draft' or a specific version_id
- Max job duration
max_job_durationtext - Defines job execution strategy. 'hours' routes to a dedicated worker tier with a concurrency of 1 and slower autoscaling , only use when jobs genuinely need >1h runtime. 'minutes' (default) uses the standard tier with higher concurrency.One of:
hours,minutes,synchronous_seconds(Synchronous seconds),background_seconds(Background seconds). - Job id
job_idtext - Can be used to control the job id.
Outputs
dataobject- Api result.
Example output
{
"data": {
"cost": 1.0,
"credits_used": [],
"errors": [],
"executionTime": 1.0,
"is_new_billing": false,
"output": {},
"status": "complete",
"step_errors": [],
"type": "credit"
}
}Start tool job
Start tool job.
Inputs
- Studio
studio_idtextRequired - Studio id.
- Project
projecttext - This can be provided to execute a studio in a project different to the one specified in the auth header.
- Params
paramsobjectRequired - Params.
- Version
versiontext - Version.
- Tool version
tool_versiontext - The version of the tool, can be 'active', 'draft' or a specific version_id
- Max job duration
max_job_durationtext - Defines job execution strategy. 'hours' routes to a dedicated worker tier with a concurrency of 1 and slower autoscaling , only use when jobs genuinely need >1h runtime. 'minutes' (default) uses the standard tier with higher concurrency.One of:
hours,minutes,synchronous_seconds(Synchronous seconds),background_seconds(Background seconds). - Job id
job_idtext - Can be used to control the job id.
Outputs
dataobject- Api result.
Example output
{
"data": {
"job_id": "example",
"project": "example",
"studio_id": "example"
}
}Trigger agent
Trigger agent.
Inputs
- Agent
agent_idtextRequired - Agent id.
- Conversation id
conversation_idtext - Conversation id.
- Metadata
metadataobject - Any additional metadata to be stored with the message. This is not sent to the agent.
- External id
external_idtext - The External ID for the conversation from the source system. By default when provided it will not create second conversations if one with the same external_id already exists.
- Params
paramsobject - Params to set in the agent before running
- Mode
modetext - If set to 'chat', this task is viewable using the chat UI on the frontendOne of:
chat.
Outputs
dataobject- Api result.
Example output
{
"data": {
"agent_id": "example",
"conversation_id": "example",
"job_info": {
"job_id": "example",
"studio_id": "example"
},
"queued_mid_run": false,
"state": "idle"
}
}Upsert agent
Upsert agent.
Inputs
- Agent
agent_idtext - Agent id.
- Name
nametext - Name.
- Description
descriptiontext - Description.
- System prompt
system_prompttext - System prompt.
- Model
modeltext - Model.One of:
openai-gpt-4o,openai-gpt-4o-2024-08-06,gpt-4o-2024-11-20,openai-gpt-4o-mini,openai-gpt-4o-mini-2024-07-18,openai-o1-latest,openai-o1-2024-12-17,openai-o3,openai-o3-2025-04-16,openai-o3-mini, and 589 more. - Temperature
temperaturenumber - Temperature of the selected model. Typically, higher temperature means more random output.
- Actions
actionslist - Actions.
- Partial update
partial_updatetrue/false - If true, updates are merged with existing agent config. If false, updates overwrite existing config.
Outputs
dataobject- Api result.
Example output
{
"data": {
"agent_id": "example"
}
}