Posthog MCP server

Read from the package source without running it. The installed server may list more.

• add_trace_to_review_queue

  Add a trace to an LLM analytics review queue for manual inspection and evaluation. This operation associates a specific trace with a review queue to enable team review workflows.

• analyze_sentiment_for_llm_outputs

  Analyze sentiment for one or more LLM outputs within a project. Performs sentiment analysis on the specified outputs at the requested analysis level (defaults to trace-level detail).

• append_segments_to_desktop_recording

  Append transcript segments to a desktop recording, supporting batched real-time streaming for continuous transcription updates.

• archive_llm_prompt_by_name

  Archive an LLM prompt by its name within a specific project. This operation creates an archived version of the prompt, preserving it for historical reference while removing it from active use.

• create_desktop_recordingwrite action

  Initiate a new desktop recording session and obtain an upload token for the Recall.ai desktop SDK to use during the recording.

• create_error_tracking_issuewrite action

  Create a new error tracking issue in a project. Requires specifying the assignee, initial status, and when the error was first detected.

• create_error_tracking_releasewrite action

  Create a new release entry for error tracking in a specified project. This registers a release version with associated metadata for tracking errors across different application versions.

• create_evaluationwrite action

  Create a new evaluation for a project to assess model outputs or behaviors. Evaluations can use LLM-based judging or Hog-based analysis and return boolean pass/fail results.

• create_llm_analytics_clustering_jobwrite action

  Create a new clustering job configuration for LLM analytics. Each team can have a maximum of 5 active clustering jobs.

• create_llm_analytics_review_queuewrite action

  Create a new review queue for LLM analytics within a specified project. Review queues organize and manage items for human review in the LLM analytics pipeline.

• create_llm_analytics_score_definitionwrite action

  Create a new LLM analytics score definition for evaluating model outputs. Score definitions are immutable after creation and define how responses will be scored using categorical, numeric, or boolean metrics.

• create_llm_promptwrite action

  Create a new LLM prompt in a project. The prompt can be defined as either a JSON object or string data and must have a unique name within the project.

• create_trace_reviewwrite action

  Create a review for an LLM trace, recording human feedback, scores, and optional reasoning. Only one active review can exist per trace and team.

• delete_clustering_jobwrite action

  Delete a clustering job configuration. Each team can maintain a maximum of 5 clustering jobs; removing one frees capacity for new configurations.

• delete_desktop_recordingwrite action

  Permanently delete a desktop meeting recording by its ID. This operation removes the recording and all associated data from the specified project.

• delete_error_tracking_assignment_rulewrite action

  Delete a specific error tracking assignment rule from a project. This removes the rule and stops it from being applied to new error assignments.

• delete_error_tracking_grouping_rulewrite action

  Permanently delete an error tracking grouping rule from a project. This removes the rule and stops it from being applied to future error grouping operations.

• delete_error_tracking_releasewrite action

  Permanently delete an error tracking release from a project. This removes the release and all associated error tracking data.

• delete_error_tracking_suppression_rulewrite action

  Permanently delete an error tracking suppression rule from a project. This removes the rule and stops it from suppressing matching errors.

• delete_error_tracking_symbol_setwrite action

  Permanently delete an error tracking symbol set from a project. This removes the symbol set and all associated symbol data used for error tracking.

• delete_llm_analytics_review_queue_itemwrite action

  Permanently delete a specific review queue item from LLM analytics. This removes the item and its associated data from the review queue.

• delete_llm_provider_keywrite action

  Permanently delete an LLM provider API key from a project. This removes the stored credentials and disables any analytics collection using this key.

• delete_review_queuewrite action

  Permanently delete a specific LLM analytics review queue from a project. This operation removes the queue and all associated data.

• delete_trace_reviewwrite action

  Permanently delete a specific trace review from an LLM analytics project. This operation removes the trace review record and cannot be undone.

• duplicate_llm_promptwrite action

  Creates a duplicate copy of an existing LLM prompt with a new unique name. The duplicated prompt retains all configuration and settings from the original.

• finish_symbol_set_bulk_uploadwrite action

  Completes a bulk symbol set upload for error tracking in a project. This finalizes the upload process after all symbol files have been transferred.

• generate_llm_evaluation_summary

  Generate an AI-powered analysis of evaluation results that identifies patterns in passing and failing evaluations and provides actionable recommendations for improving LLM response quality.

• get_clustering_job

  Retrieve a specific LLM analytics clustering job configuration by its unique identifier. Each team can maintain up to 5 clustering job configurations.

• get_conversation

  Retrieve a specific conversation by its UUID within a project. Use this to fetch detailed information about a single conversation.

• get_customer_profile_config

  Retrieve a specific customer profile configuration by its unique identifier within a project. Use this to fetch detailed settings and properties for a customer profile config.

• get_desktop_recording

  Retrieve a specific desktop meeting recording by its unique identifier within a project.

• get_error_fingerprint

  Retrieve a specific error tracking fingerprint by its unique identifier within a project. Fingerprints group similar errors together for consolidated tracking and analysis.

• get_error_tracking_assignment_rule

  Retrieve a specific error tracking assignment rule by its unique identifier. This rule defines how errors are automatically assigned within your project.

• get_error_tracking_grouping_rule

  Retrieve a specific error tracking grouping rule by its unique identifier. Grouping rules define how errors are automatically grouped and categorized within a project.

• get_error_tracking_issue

  Retrieve detailed information about a specific error tracking issue within a project, including its status, occurrences, and metadata.

• get_error_tracking_issue_activity

  Retrieve the activity history for a specific error tracking issue, including all events and changes associated with that issue.

• get_error_tracking_release

  Retrieve detailed information about a specific error tracking release for a project, including its metadata and associated error data.

• get_error_tracking_release_by_hash

  Retrieve detailed information about a specific error tracking release using its hash identifier. This operation fetches release metadata and associated error tracking data for a given project.

• get_error_tracking_suppression_rule

  Retrieve a specific error tracking suppression rule by its ID. Use this to fetch details about an existing suppression rule configured for your project.

• get_error_tracking_symbol_set

  Retrieve a specific error tracking symbol set by its unique identifier. Symbol sets contain debug information used to map obfuscated stack traces back to original source code.

• get_evaluation

  Retrieve a specific evaluation by its ID within a project. Returns detailed information about the evaluation including its configuration, results, and metadata.

• get_health_issue

  Retrieve a specific health issue by its unique identifier within a project. Returns detailed information about the health issue including its status, severity, and related metadata.

• get_llm_analytics_evaluation_config

  Retrieve the LLM analytics evaluation configuration for a specific project, which defines how evaluation metrics and criteria are applied to LLM outputs.

• get_llm_analytics_score_definition

  Retrieve a specific LLM analytics score definition by its unique identifier within a project. Use this to fetch configuration and metadata for a particular scoring metric.

• get_llm_prompt_by_name

  Retrieve a specific LLM prompt configuration by its name within a project. Use this to fetch prompt details for execution or inspection.

• get_llm_provider_key

  Retrieve a specific LLM provider key configuration by its unique identifier within a project. This returns the details of a single provider key used for LLM analytics.

• get_review_queue

  Retrieve a specific LLM analytics review queue by its unique identifier within a project.

• get_review_queue_item

  Retrieve a specific LLM analytics review queue item by its unique identifier within a project.

• get_trace_review

  Retrieve a specific trace review by its unique identifier within a project. Use this to fetch detailed information about a single LLM analytics trace review.

• list_clustering_jobs

  Retrieve a paginated list of LLM analytics clustering job configurations for a project. Each team can maintain up to 5 clustering job configurations.

• list_conversations

  Retrieve a paginated list of conversations for a specific project. Use limit and offset parameters to control pagination through the results.

• list_customer_profile_configs

  Retrieve a paginated list of customer profile configurations for a specific project. Use limit and offset parameters to control pagination through the results.

• list_dashboard_collaborators

  Retrieve the list of collaborators who have access to a specific dashboard within a project. This shows all users with whom the dashboard is shared and their access permissions.

• list_dependent_configs_for_provider_key

  Retrieve all evaluations and configurations that depend on a specific LLM provider key, along with alternative keys that can be used as replacements.

• list_desktop_recordings

  Retrieve a paginated list of desktop meeting recordings for a specific project. Use limit and offset parameters to control pagination.

• list_error_tracking_assignment_rules

  Retrieve a paginated list of error tracking assignment rules configured for a project. Assignment rules determine how errors are automatically routed to team members.

• list_error_tracking_fingerprints

  Retrieve a paginated list of error fingerprints for a specific project. Fingerprints group similar errors together to help identify recurring issues.

• list_error_tracking_grouping_rules

  Retrieve the error tracking grouping rules configured for a project. Grouping rules determine how errors are automatically grouped and deduplicated in error tracking.

• list_error_tracking_issues

  Retrieve a paginated list of error tracking issues for a specific project. Use limit and offset parameters to control pagination through the results.

• list_error_tracking_issues_activity

  Retrieve the activity log for error tracking issues within a specific project. This endpoint returns a timeline of events and changes related to tracked errors.

• list_error_tracking_releases

  Retrieve a paginated list of releases for error tracking in a specific project. Use this to view all tracked releases and their associated error data.

• list_error_tracking_spike_events

  Retrieve a paginated list of error tracking spike events for a specific project. Spike events represent sudden increases in error occurrences that warrant attention.

• list_error_tracking_suppression_rules

  Retrieve a paginated list of error tracking suppression rules configured for a specific project. Suppression rules allow you to filter out or ignore certain errors from being tracked.

• list_error_tracking_symbol_sets

  Retrieve a paginated list of symbol sets for error tracking in a specific project. Symbol sets are used to map obfuscated stack traces back to original source code.

• list_evaluations

  Retrieve a paginated list of evaluations for a specific project, with optional filtering by enabled status, evaluation IDs, and sorting capabilities.

• list_health_issues

  Retrieve a paginated list of health issues for a specific project. Health issues represent potential problems or warnings detected in the project's environment.

• list_llm_analytics_provider_keys

  Retrieve a paginated list of LLM analytics provider API keys configured for a specific project. Use this to view all provider credentials and their configurations.

• list_llm_analytics_review_queues

  Retrieve a paginated list of LLM analytics review queues for a specific project. Results can be sorted by name, creation date, or last update time.

• list_llm_analytics_score_definitions

  Retrieve a paginated list of LLM analytics score definitions for a project, with optional filtering by scorer kind and sorting capabilities.

• list_llm_prompts

  Retrieve a paginated list of LLM prompts for a specific project. Use limit and offset parameters to control pagination through the results.

• list_review_queue_items

  Retrieve a paginated list of items in LLM analytics review queues for a specific project, with optional filtering and sorting capabilities.

• list_trace_reviews

  Retrieve a paginated list of LLM trace reviews for a project, optionally filtered by scorer definition and sorted by creation or update time.

• remove_dashboard_collaboratorwrite action

  Remove a collaborator from a dashboard by their user UUID. This revokes the specified user's access to the dashboard.

• send_conversation_messagewrite action

  Send a message to a conversation or stream results from an existing conversation. Provide a message to start new conversation processing, or omit it to stream from an existing conversation.

• start_symbol_sets_bulk_uploadwrite action

  Initiate a bulk upload process for error tracking symbol sets. This operation prepares the system to receive multiple symbol set files and returns upload credentials or a session reference for subsequent file uploads.

• update_clustering_jobwrite action

  Update an existing LLM analytics clustering job configuration. Each team can maintain up to 5 clustering jobs for analyzing trace or generation-level data.

• update_clustering_job_partialwrite action

  Partially update an LLM analytics clustering job configuration. Allows modification of analysis level, event filters, and enabled status for an existing clustering job within a project.

• update_desktop_recordingwrite action

  Update metadata and processing status for a desktop meeting recording, including transcript segments, extracted tasks, and participant information.

• update_desktop_recording_partialwrite action

  Partially update a desktop recording's metadata, status, and extracted content. Use this to modify recording details, track processing status, store transcripts and tasks, or mark recordings as ready or failed.

• update_error_tracking_assignment_rulewrite action

  Update an error tracking assignment rule for a project, including its filters, ordering, assignee, and enabled/disabled status.

• update_error_tracking_assignment_rule_partialwrite action

  Partially update an error tracking assignment rule for a project. Allows modification of assignee, filters, ordering, and enabled/disabled status.

• update_error_tracking_grouping_rulewrite action

  Update an error tracking grouping rule for a project. Modify how errors are grouped together by specifying filters, ordering, and optional assignee information.

• update_error_tracking_grouping_rule_partialwrite action

  Partially update an error tracking grouping rule for a project. Allows modification of rule configuration including filters, ordering, assignee, and enabled/disabled status.

• update_error_tracking_issuewrite action

  Update an error tracking issue's status, assignment, cohort, and metadata. Modify issue details such as name, description, and linked external issues.

• update_error_tracking_issue_partialwrite action

  Partially update an error tracking issue, allowing you to modify its status, assignee, name, description, and linked external issues.

• update_error_tracking_releasewrite action

  Update an error tracking release for a project, including its version, hash identifier, and associated metadata.

• update_error_tracking_release_partialwrite action

  Partially update an error tracking release for a project. Use this to modify specific fields of an existing release without affecting other attributes.

• update_error_tracking_suppression_rulewrite action

  Update an existing error tracking suppression rule for a project. Modify filtering criteria, ordering, and sampling behavior to control which errors are suppressed.

• update_error_tracking_suppression_rule_partialwrite action

  Partially update an error tracking suppression rule for a project. Allows modification of filtering criteria, ordering, disabled status, and sampling configuration.

• update_error_tracking_symbol_setwrite action

  Update an error tracking symbol set for a project. Symbol sets are used to map obfuscated or minified code back to original source for better error tracking and debugging.

• update_evaluationwrite action

  Update an existing evaluation configuration within a project. Modify evaluation criteria, type, output settings, and associated conditions to refine how responses are assessed.

• update_evaluation_partialwrite action

  Partially update an evaluation configuration within a project. Allows modification of evaluation settings such as type, enabled status, description, and associated configurations.

• update_health_issuewrite action

  Partially update a health issue within a project, allowing you to modify specific fields such as the dismissed status without affecting other properties.

• update_llm_analytics_score_definitionwrite action

  Partially update an LLM analytics score definition within a project. Allows modification of scorer metadata such as description while preserving other configuration details.

• update_llm_provider_keywrite action

  Update an LLM provider API key configuration for analytics tracking. Modify the provider type and display name for an existing key credential.

• update_llm_provider_key_partialwrite action

  Partially update an LLM provider API key configuration for a specific project. Allows modification of the provider type or other key attributes without replacing the entire configuration.

• update_review_queuewrite action

  Partially update an LLM analytics review queue configuration within a specific project. Use this to modify queue settings, filters, or other properties without replacing the entire queue.

• update_review_queue_itemwrite action

  Partially update a review queue item by reassigning it to a different review queue or modifying its metadata. Use this to move pending traces between queues or update queue ownership.

• update_trace_reviewwrite action

  Partially update a trace review with new scores, comments, or queue context. Clears associated pending queue items upon save.

• validate_llm_provider_key

  Validate an LLM provider API key to ensure it is correctly configured and has proper permissions. This operation tests the key's authenticity and functionality with the specified provider.

• Code scan5 source files scanned25/25
• –Live reliabilityno gateway calls yet and no remote to proben/a
• –Tool poisoningtools not inspected (local package is not executed); not countedn/a
• Auth qualitylocal package, no credentials required12/15
• Maintenancelast push 138 days ago8/15
• Maintainer identityregistry namespace matches repository owner7/10

Safety score

80/100

higher than 60% of 48,255 servers

Tools

100

the typical server has 7

Tools that change things

49%

typical: 0%. Fewer write tools means less an agent can break

GitHub stars

31

more than 92% of 28,688 servers

Posthog: common questions

Is Posthog MCP server safe?

Mostly: it is graded B (80/100). Read the Posthog safety report

How do I install Posthog?

It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.

Does Posthog need an API key?

Not as far as the registry entry and our scan can tell: no credentials are declared or required.

Is Posthog maintained?

The last commit was 176 days ago (2026-05-12). The latest release is v1.0.5.

How does Posthog compare with other MCP servers?

safety score 80/100 (higher than 60% of 48,255 servers); tools 100 (the typical server has 7); tools that change things 49% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.