Integration

CustomGPT.ai integration

CustomGPT.ai builds retrieval-augmented chatbots (agents) trained on your own content. Create and configure agents, add sitemap or file sources and manage their documents, chat with an agent through conversations or an OpenAI-compatible completion route, read citations, claims and trust scores, run agent tasks, manage tags, licenses, child agents, personas and custom actions, and read traffic, query, conversation and lead reports plus account limits. Payment methods and invoice PDF downloads are not exposed, and replies are returned whole (streaming is off). Authentication is a CustomGPT API token sent as a Bearer token; create it on the Developers page of the dashboard.

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in CustomGPT.ai.Built from the 117 CustomGPT.ai tools below, on the connection you add once.

What you can do with CustomGPT.ai

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Label, 8 operations

  • assign_tag_usersAssign tag usersPUTSets which users belong to each tag. The list you send becomes the truth for those tags: users left out are removed from them. Requires a plan with tags.
  • create_tagsCreate tagsPOSTCreates one or more new tags for document access control. Accepts an array of tag names, allowing bulk creation.
  • delete_tagDelete tagDELETEDeletes a tag and all its user and document assignments. Documents left with no tag become visible to everyone.
  • get_document_tagsGet document tagsGETReturns all labels assigned to a document.
  • get_tagGet tagGETRetrieves a specific tag, including its name and the list of users assigned to it.
  • list_tagsList tagsGETReturns all tags configured for an agent, along with the users assigned to each tag. Use this to review your access control configuration or build a tag management interface.
  • rename_tagRename tagPUTUpdates the name of an existing tag. User and document assignments are preserved.
  • replace_document_tagsReplace document tagsPUTReplaces the tags of a document with the label_ids you send (an empty list removes all, making it visible to everyone). Maximum 50 tags per document.

Message, 7 operations

  • get_messageGet messageGETReturns the complete details for a single message, including the user's prompt, the agent's response, timestamps, citations, and any attached metadata.
  • get_task_messageGet task messageGETReturns the complete details for a single message, including the user's prompt, the agent's response, timestamps, citations, and any attached metadata.
  • list_messagesList messagesGETReturns all messages in a conversation, including user prompts, agent responses, timestamps, and citations. Use this to display conversation history or analyze past interactions.
  • list_task_messagesList task messagesGETReturns all messages in a task, including user prompts, agent responses, timestamps, and citations. Use this to display task history or analyze past interactions.
  • send_messageSend messagePOSTSends a prompt in a conversation and returns the agent's full reply with citations. Use a session_id from create_conversation. Each call consumes one query of the plan. Replies are returned whole, not streamed.
  • send_task_messageSend task messagePOSTSends a follow-up prompt to a task that is not processing (409 otherwise). Returns 202 at once; call get_task to follow progress and read the result.
  • update_webpage_context_messagesUpdate webpage context messagesPATCHUpdates `auto_message` and/or `display_message` for one or more webpage context entries. Requires the Webpage Awareness tool to be enabled on the agent.

Conversation, 5 operations

  • create_conversationCreate conversationPOSTCreates a new conversation session for an agent. Returns a session ID that you'll use to send messages. You can optionally provide a name to help identify the conversation later.
  • delete_conversationDelete conversationDELETEPermanently deletes a conversation and all its messages. Use this to remove sensitive conversations or clean up test data. Warning: This action cannot be undone.
  • get_conversation_reportGet conversation reportGETReturns conversation metrics, including total conversations, average duration, messages per conversation, and completion rates.
  • list_conversationsList conversationsGETReturns all conversations for an agent. Each conversation represents a chat session with a user and contains the full message history.
  • update_conversationUpdate conversationPUTUpdates conversation metadata, such as the display name. Use this to rename conversations for easier organization and identification.

External apis, 5 operations

  • create_external_apiCreate external apiPOSTConnects an API to an agent from the public HTTPS URL of its OpenAPI schema. The agent calls it only with Plan & Act and Smart Tasks on. At most 5 can be active; a sixth is created disabled.
  • delete_external_apiDelete external apiDELETEDeletes a custom External API from an agent, by the id from list_external_apis.
  • get_external_apiGet external apiGETReturns one custom External API connected to an agent, by the id from list_external_apis.
  • list_external_apisList external apisGETLists the custom External APIs (OpenAPI-described APIs the agent may call) connected to an agent, and whether each is available.
  • update_external_apiUpdate external apiPUTRenames it, replaces its API key, or enables/disables it. At most 5 External APIs can be active at once.

License, 5 operations

  • create_licenseCreate licensePOSTCreates a new access license for an agent with specified permissions and optional expiration. Use this to grant controlled access to external users or systems.
  • delete_licenseDelete licenseDELETERevokes and permanently removes an access license. Any users or systems using this license will immediately lose access to the agent.
  • get_licenseGet licenseGETReturns the full details for a specific license, including permissions, usage statistics, status, and expiration date. Use this to audit or troubleshoot access issues.
  • list_licensesList licensesGETReturns all access licenses configured for an agent. Licenses control who can interact with the agent and what permissions they have. Use this to manage access distribution.
  • update_licenseUpdate licensePUTUpdates a license's settings, such as permissions, expiration date, or active status. Use this to modify access without creating a new license.

Project, 5 operations

  • create_agentCreate agentPOSTCreates an agent from a sitemap or website URL (sitemap_path) or one uploaded file (file). Indexing starts at once and can take a while; check get_agent or list_documents for progress.
  • delete_agentDelete agentDELETEPermanently deletes an agent and all associated data, including its knowledge base, conversations, settings, and analytics. This action cannot be undone.
  • get_agentGet agentGETReturns the full configuration and current status for a specific agent. Use this to check processing status, view settings, or retrieve metadata about the agent.
  • list_agentsList agentsGETReturns all agents in your account, sorted by creation date with the most recent first.
  • update_agentUpdate agentPOSTUpdates an agent's name or settings, or adds content from a sitemap_path or an uploaded file. Send only the fields to change; leave a field out to keep the current setting.

Task, 5 operations

  • create_taskCreate taskPOSTStarts a Plan & Act task with its first prompt and returns its id at once (202). The task runs in the background: call get_task with data.id until status is no longer processing.
  • delete_taskDelete taskDELETEPermanently deletes a task and all its messages. Use this to remove sensitive tasks or clean up test data. Warning: This action cannot be undone.
  • get_taskGet taskGETReturns a task's state: processing (with events since the last poll), completed with the result, or awaiting_tool_results with the tool calls to answer through submit_task_tool_outputs.
  • list_tasksList tasksGETReturns all Plan & Act mode tasks for an agent, paginated. Tasks are ordered by most recent activity.
  • update_taskUpdate taskPUTUpdates task metadata, such as the display name. Use this to rename tasks for easier organization and identification.

Child, 4 operations

  • assign_child_agentsAssign child agentsPOSTAssigns one or more agents as children of this Multi-Agent. Enforces plan limits, team ownership, and prevents assigning Multi-Agent projects as children.
  • list_child_agentsList child agentsGETReturns all agents currently assigned to this Multi-Agent, including their name and generation flag.
  • unassign_child_agentUnassign child agentDELETERemoves a single child agent from this Multi-Agent. Returns 422 if this is the last remaining child, or 404 if the agent is not currently assigned.
  • unassign_child_agentsUnassign child agentsDELETERemoves multiple child agents from this Multi-Agent in a single request. Returns 422 if the operation would leave no children, or if any specified agent is not currently assigned.

Page, 4 operations

  • delete_documentDelete documentDELETEPermanently removes a document from an agent's knowledge base. The agent will no longer reference this content when answering questions.
  • delete_webpage_contextsDelete webpage contextsDELETEDeletes webpage context entries for an agent. Pass `page_ids` to delete specific pages, or omit the body to delete all pages for the agent.
  • list_documentsList documentsGETReturns all indexed documents in an agent's knowledge base.
  • list_webpage_contextsList webpage contextsGETReturns all webpage context entries for an agent, with optional URL filtering and pagination. Requires the Webpage Awareness tool to be enabled on the agent.

Role, 4 operations

  • create_team_roleCreate team rolePOSTCreates a new custom role for the team with the specified permissions. Role names cannot match reserved role names (Owner, Administrator, Member).
  • delete_team_roleDelete team roleDELETEDeletes a custom role belonging to this team.
  • list_team_rolesList team rolesGETReturns all roles that can be assigned to members within this team. Includes both global assignable system roles (Administrator, Member, Chat-only) and the team's own custom roles.
  • update_team_roleUpdate team rolePUTUpdates an existing custom role's name, description, and permissions. Only roles belonging to this team can be updated.

Source, 4 operations

  • create_sourceCreate sourcePOSTAdds a sitemap or website URL (sitemap_path) or one uploaded file (file) as a new source of an agent. Indexing can take several minutes; check list_sources or list_documents.
  • delete_sourceDelete sourceDELETERemoves a source and every document it brought into the agent's knowledge base. To remove a single document only, use delete_document.
  • list_sourcesList sourcesGETReturns all data sources connected to an agent. Each source represents an origin for documents - such as a sitemap, Google Drive folder, SharePoint site, or uploaded files.
  • update_sourceUpdate sourcePUTUpdates configuration for a data source, such as auto-sync frequency, crawl depth, or file filters. Use this to fine-tune how the source is indexed and kept up to date.

Custom, 3 operations

  • create_custom_actionCreate custom actionPOSTAdds a custom action to an agent. Only the MCP connector type is supported; settings carry the connector configuration.
  • delete_custom_actionDelete custom actionDELETEDeletes a custom MCP action from an agent, identified by its tool key from list_agent_actions.
  • update_custom_actionUpdate custom actionPUTUpdates a custom MCP action of an agent (name, description, settings) or enables or disables it. Leave a field out to keep the current value.

Export, 3 operations

  • export_conversationExport conversationGETExports a conversation's full history as a file (json, excel, csv or pdf, set by format) and returns it as a stored file.
  • export_conversation_leadsExport conversation leadsGETExports every lead captured in one conversation as a file (json, excel, csv or pdf, set by format) and returns it as a stored file.
  • export_taskExport taskGETExports a task's full history (prompts and responses) as a file (json, excel, csv or pdf, set by format) and returns it as a stored file.

Invitation, 3 operations

  • cancel_team_invitationCancel team invitationDELETECancels a pending team invitation. The invited person will no longer be able to join via this invitation link.
  • invite_team_memberInvite team memberPOSTSends an invitation email to the address given; the person joins the team once they accept.
  • list_team_invitationsList team invitationsGETReturns a paginated list of all pending invitations for the specified team.

Member, 3 operations

  • list_team_membersList team membersGETReturns a paginated list of all members belonging to the specified team, including their assigned roles.
  • remove_team_memberRemove team memberDELETERemoves a member from the team. Only the team owner or a member with the appropriate permission can perform this action.
  • update_team_member_roleUpdate team member rolePUTChanges the role assigned to a team member. Only the team owner or a member with the appropriate permission can perform this action.

Settings, 3 operations

  • get_agent_settingsGet agent settingsGETReturns the complete configuration for an agent, including persona instructions, response behavior, deployment options, and feature flags.
  • update_agent_settingsUpdate agent settingsPOSTUpdates an agent's settings (persona, model, citations, messages, appearance, retention). Send only the fields to change; the others keep their value.
  • update_builtin_action_settingsUpdate builtin action settingsPUTUpdates the settings of one built-in action of an agent, identified by its key from list_agent_actions.

Sub persona, 3 operations

  • delete_sub_persona_overrideDelete sub persona overrideDELETEDeletes the project-level override for one sub-persona slug. Generated persona-version rows are not modified.
  • list_sub_personasList sub personasGETReturns generated sub-personas for one persona version and project-level overrides separately.
  • update_sub_persona_overrideUpdate sub persona overridePATCHCreates or updates the project-level override for one sub-persona slug. Generated persona-version rows are not modified.

User, 3 operations

  • get_user_profileGet user profileGETReturns profile information for the currently authenticated user, including account details and settings.
  • remove_tag_usersRemove tag usersDELETERemoves specific users from one or more tags without affecting other users' assignments. Use this when offboarding users or revoking access to specific content.
  • update_user_profileUpdate user profilePOSTUpdates the profile of the user who owns the API token. Leave a field out to keep its current value.

Deployment, 2 operations

  • get_deployment_settingsGet deployment settingsGETReturns all deployment section configurations (embed, live_chat, ai_assistant, instant_viewer, sge) for an agent. Missing settings are returned with their default values.
  • update_deployment_settingsUpdate deployment settingsPATCHUpdates one deployment section of an agent (embed, live_chat, ai_assistant, instant_viewer or sge). Send only the fields to change; the others keep their value.

File, 2 operations

  • delete_task_message_fileDelete task message fileDELETEDeletes an agent-generated artifact from a task message.
  • list_task_message_filesList task message filesGETReturns all files associated with a specific task message - both user-uploaded sources and AI-generated artifacts.

Metadata, 2 operations

  • get_document_metadataGet document metadataGETReturns metadata for a document, including its title, source URL, word count, and any custom metadata fields.
  • update_document_metadataUpdate document metadataPUTUpdates custom metadata fields for a document. Use this to add tags, categories, or other organizational information that helps manage your knowledge base.

Persona, 2 operations

  • get_persona_versionGet persona versionGETReturns the full text and metadata for a specific persona version. Use this to compare versions or review what instructions were active at a particular time.
  • list_persona_versionsList persona versionsGETReturns the version history of an agent's persona.

Team, 2 operations

  • list_teamsList teamsGETReturns a paginated list of all teams the authenticated user is a member of.
  • rename_teamRename teamPUTUpdates the name of a team. Only the team owner or a member with the update permission can perform this action.

Version, 2 operations

  • get_skill_versionGet skill versionGETReturns the full text and metadata of one version of a skill.
  • list_skill_versionsList skill versionsGETReturns the version history of one skill, newest first.

Action, 1 operation

  • list_agent_actionsList agent actionsGETLists the actions (built-in tools and custom MCP actions) available to an agent, with their keys, names and whether each is enabled.

Activate, 1 operation

  • restore_skill_versionRestore skill versionPUTMakes one version of a skill the live skill body again.

Analysis, 1 operation

  • get_analytics_chartGet analytics chartGETReturns time-series data formatted for charts, with daily, weekly, or monthly breakdowns of key metrics.

Citation, 1 operation

  • get_citationGet citationGETReturns metadata for a specific citation, including the source document's title, description, URL, and preview information.

Claim, 1 operation

  • list_message_claimsList message claimsGETReturns the factual claims extracted from an agent's response. Each claim is a discrete statement that can be independently verified against the source documents.

Collated, 1 operation

  • get_collated_leadGet collated leadGETReturns one merged lead built from all lead rows in the conversation.

Completion, 1 operation

  • create_chat_completionCreate chat completionPOSTSends OpenAI-style messages to an agent and returns an OpenAI-format chat completion answered from the agent's content. Each call consumes one query of the plan. Streaming is off.

Feedback, 1 operation

  • update_message_feedbackUpdate message feedbackPUTRecords user feedback (thumbs up/down) for a message. Use this to collect quality signals that help identify which responses are helpful and which need improvement.

Instant sync, 1 operation

  • sync_sourceSync sourcePUTRe-syncs a source now, refreshing all its documents, instead of waiting for its scheduled refresh.

Intelligence, 1 operation

  • get_customer_intelligenceGet customer intelligenceGETReturns AI-analyzed insights about your users, including common intents, emotional sentiment, frequently discussed topics, and emerging trends.

Invoice, 1 operation

  • list_invoicesList invoicesGETReturns paginated invoices for your account. Invoice records are stored locally and do not require a Stripe API call. Use the `page` and `limit` query parameters to paginate.

Lead, 1 operation

  • list_leadsList leadsGETLists the leads captured in an agent's conversations (emails, names, custom fields), paginated, optionally for one conversation (session_id) or a date range.

Openapi, 1 operation

  • replace_external_api_schemaReplace external api schemaPOSTReplaces the OpenAPI schema of a custom External API with the one at openapi_url (public HTTPS, JSON or YAML).

Persona activate, 1 operation

  • activate_persona_versionActivate persona versionPUTMakes an earlier persona version the active persona again. A new version entry is added; newer versions are kept.

Plan, 1 operation

  • get_billing_planGet billing planGETReturns the active subscription plan for your account, including billing cycle dates, trial status, plan options, and active addons.

Query, 1 operation

  • get_query_reportGet query reportGETReturns aggregate query metrics, including total query volume, success rates, average response times, and failure reasons.

Reindex, 1 operation

  • reindex_documentReindex documentPOSTRe-crawls and re-indexes one URL-based document so the agent uses its latest content. Uploaded files cannot be reindexed: delete and upload them again.

Replicate, 1 operation

  • replicate_agentReplicate agentPOSTCreates a complete copy of an existing agent, including its knowledge base, persona, and settings.

Resend, 1 operation

  • resend_team_invitationResend team invitationPOSTCancels a pending invitation and sends a new invitation email to the same address with the same role.

Search team member, 1 operation

  • search_team_membersSearch team membersGETSearches for team members by email address or user ID. Use this to find users when assigning permissions or managing team access. Notes: Requires Owner or Admin role.

Stats, 1 operation

  • get_agent_statsGet agent statsGETReturns usage metrics and performance statistics for an agent, including total conversations, query counts, document statistics, and processing information.

Stop, 1 operation

  • stop_taskStop taskPOSTStops the response a task is generating. Work already done is saved and billed, and the plan is kept so the next message can continue it. Follow with get_task.

Switch, 1 operation

  • switch_teamSwitch teamPOSTSwitches the active team of the API token's user: every later call then works on that team's agents. Team ids come from list_teams.

Tool output, 1 operation

  • submit_task_tool_outputsSubmit task tool outputsPOSTResumes a task paused on awaiting_tool_results with the results of the function calls the agent asked for. Answer every pending tool_call_id in one call. Returns 202; follow with get_task.

Traffic, 1 operation

  • get_traffic_reportGet traffic reportGETReturns a breakdown of conversation volume by deployment channel (e.g. web widget, API, Slack, WhatsApp) over the requested time window.

Trust score, 1 operation

  • get_message_trust_scoreGet message trust scoreGETReturns a verification trust score for a message, calculated by checking how well the agent's claims are supported by the source documents.

Usage, 1 operation

  • get_usage_limitsGet usage limitsGETReturns your current usage and plan limits for agents, storage (words), and monthly queries.

Verify, 1 operation

  • verify_messageVerify messagePOSTTriggers a fact-checking verification process for a message's claims.

Use CustomGPT.ai from Claude Code, Cursor or Codex

Connect CustomGPT.ai in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its CustomGPT.ai operations become tools the client can call, next to every other integration you connected.

Claude Code
claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"
Codex (~/.codex/config.toml)
[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"
Cursor, Claude Desktop and other JSON configs
{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect CustomGPT.ai

  1. Create your workspace

    Sign up free. Nothing to install.

  2. Connect CustomGPT.ai once

    Add your CustomGPT.ai connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.

  3. Describe the job in chat

    The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.

CustomGPT.ai integration FAQ

What can you do with the CustomGPT.ai integration?

LiveContext exposes 117 operations on CustomGPT.ai, covering label, message, conversation and external apis.

Popular integrations

Browse the whole catalog