Honeyhive

Tool to add datapoints to a dataset. Use when you need to append multiple entries with specified input, ground truth, and history mappings.

Untrusted display-only summary

Tool to retrieve experiment comparison between two evaluation runs. Use when you need to analyze the differences in metrics, datapoints, and events between two runs.

Untrusted display-only summary

Tool to compare events between two experiment runs side-by-side. Use when analyzing differences in model behavior, performance metrics, or outputs between evaluation runs. Returns matched event pairs with their respective data from both runs for comparison.

Untrusted display-only summary

Tool to create multiple datapoints in a single batch operation. Use when you need to bulk-import events into a dataset or create many datapoints at once. Supports filtering by date range, event IDs, or custom criteria. Efficient for migrating large numbers of events to evaluation datasets.

Untrusted display-only summary

Tool to create multiple model events in a single request. Use when you need to log a batch of event interactions to HoneyHive.

Untrusted display-only summary

Tool to log a batch of external API calls as tool events. Use when you need to record multiple tool events in one request—use after gathering all event data.

Untrusted display-only summary

Creates a new configuration in HoneyHive for managing LLM or pipeline settings. Use this to define reusable configurations with specific models, prompts, and parameters that can be deployed across different environments (dev, staging, prod). Configurations enable version control and environment-specific management of…

Untrusted display-only summary

Tool to create a new datapoint with input-output pairs. Use when you need to add a single datapoint with inputs, ground truth, conversation history, and metadata.

Untrusted display-only summary

Tool to create a dataset. Use when you need to initialize a new dataset within a project.

Untrusted display-only summary

Tool to create a new event in HoneyHive to track execution of different parts of your application. Use when you need to log a model call, tool execution, or chain step. Events can be grouped into sessions and nested hierarchically using parent_id and children_ids.

Untrusted display-only summary

Tool to create a new metric in HoneyHive. Use when you need to define how to evaluate model outputs, whether through code (PYTHON), AI evaluation (LLM), human review (HUMAN), or combining multiple metrics (COMPOSITE). Important: LLM metrics require both model_provider and model_name to be specified.

Untrusted display-only summary

Tool to create a new model event to log LLM call data. Use when you need to track a single model interaction including messages, responses, usage, and metadata.

Untrusted display-only summary

Creates a new tool definition in a HoneyHive project. Use this to register functions or plugins that can be invoked and tracked within HoneyHive. Tools are defined with a JSON Schema for their parameters, allowing HoneyHive to validate inputs and track tool usage in your AI workflows. Tool names must be unique within…

Untrusted display-only summary

Tool to delete a specific datapoint by its ID. Use when you need to remove a datapoint from HoneyHive after confirming its identifier.

Untrusted display-only summary

Tool to delete a dataset by ID. Use when you need to remove a dataset after confirming its ID.

Untrusted display-only summary

Tool to update an evaluation run's status and metadata. Use to mark a run as completed after finishing evaluations, or update run properties like name, metadata, configuration, and associated event/datapoint IDs.

Untrusted display-only summary

Tool to retrieve a list of configurations. Use when you need to fetch all configurations for a specific project before making changes.

Untrusted display-only summary

Retrieve datasets from HoneyHive for a specified project. Use this tool when you need to: - List all datasets within a project - Find datasets by type (evaluation or fine-tuning) - Retrieve a specific dataset by its ID Returns dataset details including name, description, datapoints count, type, and timestamps.

Untrusted display-only summary

Tool to query events with filters and projections from HoneyHive. Use this action when you need to retrieve events with lightweight filtering (limit 1000 results). For bulk exports or more complex queries, use the Retrieve Events action instead. Supports filtering by date range, event properties, and field projections.

Untrusted display-only summary

Tool to retrieve the complete tree of nested events for a specific session. Use when you need to analyze all events (model calls, tool calls, chains) that occurred within a session, including their hierarchical relationships, inputs, outputs, metrics, and costs. Returns a tree structure with recursive children.

Untrusted display-only summary

Tool to retrieve charting and analytics data for events over time. Use when you need aggregated metrics (duration, cost, token usage) grouped by time buckets or fields. Supports percentile analysis (p50, p95, p99) for latency monitoring and custom filters for targeted analytics.

Untrusted display-only summary

Retrieves all metrics associated with a HoneyHive project. Returns a list of metrics including their configuration (name, type, description, thresholds, evaluator details) and metadata (creation/update timestamps, sampling settings). Use this tool when you need to: - List all metrics configured for a project - Get met…

Untrusted display-only summary

Tool to retrieve all projects in the HoneyHive account. Use when you need to list available projects, get project IDs for use in other API calls, or search for a specific project by name.

Untrusted display-only summary

Tool to get details of an evaluation run by its UUID. Use when you need to check the status, configuration, results, or metadata of a specific evaluation run.

Untrusted display-only summary

Tool to get event metrics for an experiment run. Use when you need to retrieve metrics computed on events within a specific experiment run. Returns an array of event objects with their associated metrics, which can be filtered by date range or custom filters.

Untrusted display-only summary

Tool to retrieve a list of evaluation runs from HoneyHive. Use when you need to: - List all evaluation runs for analysis - Find runs by status, name, or dataset - Get specific runs by their IDs - Paginate through large sets of evaluation runs Returns evaluation details including status, results, configuration, and tim…

Untrusted display-only summary

Tool to retrieve the schema for experiment runs in HoneyHive. Use when you need to understand available fields, datasets, and mappings for experiment runs.

Untrusted display-only summary

Retrieve a complete session tree by session ID from HoneyHive. Use this tool to fetch the full session hierarchy including all nested events (model calls, tool calls, chains) with their inputs, outputs, durations, and metadata. Returns a recursive tree structure with aggregated metrics. Prerequisites: You need a valid…

Untrusted display-only summary

Tool to list all available Honeyhive tools. Use when you need to discover which functions or plugins are registered for use.

Untrusted display-only summary

Retrieve a specific datapoint by its ID from HoneyHive. Use this tool when you need the full details of a single datapoint, including its inputs, ground truth, conversation history, linked datasets, and metadata. Prerequisites: You need a valid datapoint ID. Get datapoint IDs from: - HONEYHIVE_RETRIEVE_DATAPOINTS (lis…

Untrusted display-only summary