Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.honeycomb.io/llms.txt

Use this file to discover all available pages before exploring further.

Last updated: March 16, 2026

​
Overview

Honeycomb provides this notice to describe our approach to developing artificial intelligence (AI) features and to answer frequently asked questions about Honeycomb Intelligence, a suite of AI features available through our service.

​
Our AI principles

We believe AI can meaningfully improve our services and customer experience. As we develop and deploy AI-based features, we are committed to the following principles:
  • Use AI where it makes sense: We develop AI features to enhance our products and services in places where AI can uniquely benefit them.
  • Be transparent: We scope AI features to their purpose and disclose their limitations. We do not claim capabilities that are impossible or unreasonable to perform. We make it clear when you are interacting with an AI feature and how it is being used.
  • Ensure fairness and inclusivity: We design AI features to avoid bias and discrimination, and to be useful and accessible to all.
  • Maintain reliability and safety: We design AI features to function reliably and safely. We monitor and address unreliable behaviors when they arise, including potentially removing a feature if it is deemed too problematic.
  • Protect privacy and security: We design AI features to meet the same privacy and security standards as our other product functionality, so you can trust us with your data.
  • Be accountable: We monitor AI features on an ongoing basis to ensure goals are met. We track and remediate issues when they arise.

​
Honeycomb Intelligence features

The following table describes the current Honeycomb Intelligence features, the model providers each feature uses, and how each feature interacts with your data.
FeatureDescriptionModel ProvidersData Interaction
Honeycomb CanvasAI-guided workspace inside Honeycomb that combines an AI assistant with an interactive notebook for visualizing query results and traces.OpenAI, AWS BedrockUses user-provided text, dataset/environment schema information, and sample telemetry values to create, read, and update Honeycomb queries and any entity within Honeycomb.
Honeycomb MCPInterface for your Honeycomb telemetry in any client application that connects to the Honeycomb MCP.Client-dependent/user-controlled model providerUses user-provided text, dataset/environment schema information, and sample telemetry values to create, read, and update Honeycomb queries and any entity within Honeycomb.
Query AssistantTextual interface that helps users create runnable Honeycomb queries.OpenAI, AWS BedrockUses user input, dataset/environment schema information, and sample telemetry values to produce runnable Honeycomb queries.
AI Assisted Calculated FieldsText-to-expression UI that helps users create valid Calculated Field expressions.OpenAI, AWS BedrockUses user input, dataset/environment schema information, and sample telemetry values to produce valid Calculated Fields.
BubbleUp InsightsSurfaces a plain-language summary and ranked list of fields that differ most between a BubbleUp selection and the baseline.AWS BedrockUses query results, dataset/environment schema information, and sample telemetry values to compare a BubbleUp selection against the baseline and produce a plain-language summary and ranked table of fields with insights and severity ratings.

​
FAQ

​
Is the use of Honeycomb Intelligence optional?

Yes. Team Owners can enable or disable Honeycomb Intelligence at the Team level at any time. To learn how to enable or disable Honeycomb Intelligence for your Team, visit Manage Team Behavior.

​
Are any AI features activated even if I have Honeycomb Intelligence disabled?

No. While some AI features may activate passively to surface insights proactively, all AI features are governed by the Team-level AI settings.

​
Are there limits on use of Honeycomb Intelligence?

Not at the moment. We may place limits in the future for some Honeycomb Intelligence features to mitigate costs or prevent misuse and abuse. These limits may change over time.

​
Do Honeycomb Intelligence features use generative AI exclusively?

No. Honeycomb also uses statistical and deterministic models to power some features, such as Anomaly Detection.

​
Will the data I input to Honeycomb Intelligence be used to train machine learning models?

Honeycomb does not use any AI model providers that train foundation models based on input. We may fine-tune pre-trained models to provide a better product offering. Other machine learning systems may require training a different kind of machine learning model on a per-Team basis, or performing a fit operation (training a model on your Team’s data specifically) for a statistical model.

​
Are any Honeycomb Intelligence features able to process protected health information or other sensitive data?

Yes. Teams with a Business Associate Agreement (BAA) are eligible for Honeycomb Intelligence. For more information, contact your Account Manager and review our Supplemental Terms.

​
Do all Honeycomb Intelligence model providers process personal data as a subprocessor?

In some cases, we use offline models, where the underlying model provider does not process or otherwise have access to any input data. In other cases, we may use online models, where the AI model provider may serve as a subprocessor. Where an AI model provider used by Honeycomb may receive personal data on a subprocessor basis, Honeycomb adds that provider to its subprocessor list.

​
What offline models are you using?

The offline models we use are self-hosted through AWS Bedrock, so the underlying model provider does not have access to your data.

​
Does Honeycomb Intelligence use multiple models at the same time?

Yes. We regularly test newly released models from our model providers to evaluate their efficacy. A given Honeycomb Intelligence feature may call multiple different models, sometimes from different model providers.

​
Does Honeycomb Intelligence support multi-modality (image, audio) inputs and outputs?

Not currently. We may add support for multi-modality inputs and outputs in the future.