Arize Phoenix
AI & LLMsAI Observability & Evaluation
Release History
View all versions →@arizeai/phoenix-mcp@4.3.4This patch release updates dependencies, specifically for the @arizeai/phoenix-client package to version 7.6.0.
@arizeai/phoenix-cli@1.16.2This patch release updates dependencies, specifically for the @arizeai/phoenix-client package to version 7.6.0.
@arizeai/phoenix-client@7.6.02 featuresIntroduced a new `deletePrompt` helper for managing prompts and added a `PromptIdentifier` type for improved selector handling. The deletion process now cascades to all prompt versions.
arize-phoenix-v20.4.0Breaking12 fixes10 featuresThis release introduces new features like in-process Phoenix MCP toolset for agents, retrieval relevance evaluator, and AI Query for trace filter DSL. It also includes several bug fixes and dependency updates.
arize-phoenix-sqlean-v0.1.11 fixThis release bumps the bundled sqlean version for phoenix-sqlean to 0.28.4+10a13f9, addressing a bug fix.
arize-phoenix-evals-v3.5.11 fixThis release includes a bug fix that upgrades the anthropic SDK to version 1.
arize-phoenix-client-v3.3.01 featureThis release introduces a new feature enabling phoenix-gql mutations by default in manual mode, requiring approval. This change aims to streamline mutation workflows.
arize-phoenix-evals-v3.5.01 featureThis release introduces a new retrieval relevance evaluator for evals, enhancing the evaluation capabilities of the system.
arize-phoenix-v20.3.01 featureIntroduced an expression filter DSL for traces, enhancing filtering capabilities. This release focuses on improving trace analysis with new filtering options.
arize-phoenix-v20.2.11 fixThis release includes a bug fix for cost calculation by updating built-in model token prices.
arize-phoenix-client-v3.2.01 featureThis release introduces a new POST /traces/transfer endpoint to the server, enhancing trace management capabilities.
arize-phoenix-v20.2.01 fix1 featureThis release introduces a machine-readable approval decision for agents on gated PXI tools and updates built-in model token prices. Documentation has also been improved with the addition of Phoenix release notes and a reference for span & session filter expressions.
arize-phoenix-v20.1.02 fixes3 featuresThis release introduces JWT client assertion support for OAuth2 login and adds a new /traces/transfer endpoint. It also includes bug fixes for session ID collisions and OpenAI model routing.
arize-phoenix-client-v3.1.01 featureThis release introduces agent session persistence. It also includes miscellaneous chores related to the release process.
arize-phoenix-v20.0.04 fixes4 featuresThis release introduces semantic coloring for project metric tooltips and agent session persistence. It also includes several bug fixes and dependency updates, including the arize-phoenix-client to version 3.0.0.
arize-phoenix-client-v3.0.0Breaking5 fixes15 featuresThis release introduces significant updates including server-side PXI tracing, new REST endpoints for prompt metadata and experiment tags, and an OpenAI-compatible completions proxy. It also includes a breaking change in the client's formatter for generative AI.
arize-phoenix-v19.21.07 fixes2 featuresThis release introduces new features for tagging experiments and filtering sessions with AI queries. It also includes several bug fixes related to dependencies, error handling, and UI responsiveness.
arize-phoenix-otel-v0.17.12 fixesThis release includes bug fixes for the phoenix-otel component, specifically addressing README content and processor replacement logic. Documentation has also been updated across various features.
arize-phoenix-v19.20.02 fixes2 featuresThis release introduces new features for session turn copying and dataset split management via REST endpoints. It also includes bug fixes for project navigation and dependency updates.
arize-phoenix-v19.19.16 fixesThis release includes several bug fixes, performance improvements, and documentation updates. Key fixes involve handling OTLP paths, database operations, and resource leaks. Performance enhancements focus on SQLite connection pooling and UI rendering.
arize-phoenix-otel-v0.17.02 fixes1 featureThis release introduces support for phoenix environment files and includes several bug fixes for the phoenix-otel component. Documentation has also been updated across various features.
arize-phoenix-evals-v3.4.01 featureThis release introduces a new hallucination evaluator for evals, enhancing the model evaluation capabilities.
arize-phoenix-v19.19.01 fix1 featureThis release introduces UI improvements for LLM spans and a database-related bug fix by replacing sqlean.py with arize-phoenix-sqlean.
arize-phoenix-v19.18.02 fixes2 featuresThis release introduces a new REST endpoint for prompt metadata updates and an expression filter DSL for sessions. It also includes bug fixes for cost tracking and UI rendering.
arize-phoenix-sqlean-v0.1.01 featureThis release introduces the 'phoenix-sqlean' feature, which vendors nalgeon/sqlean.py as arize-phoenix-sqlean.
Common Errors
ChatCompletionSubscriptionError2 reportsThe "ChatCompletionSubscriptionError" in arize-phoenix usually arises from insufficient OpenAI account credits or an expired/inactive subscription. To fix it, verify your OpenAI account has sufficient credits or that your paid subscription is active and properly linked to your Phoenix environment, then restart any relevant Phoenix processes. If using API keys, ensure they are valid and have appropriate permissions.
AssertionError2 reportsAssertionErrors in Arize Phoenix Anthropic streaming usually stem from unhandled event types within the `ParsedMessage` structure, specifically `ParsedMessageStopEvent` or `ParsedContentBlockStopEvent`. To resolve this, ensure your code includes explicit handling for these stop events, likely by adding new cases to your event-processing logic that gracefully exits or modifies the data stream. This might involve updating a switch statement or adding new conditional branches to properly manage these specific stop conditions.
DatatypeMismatchError1 reportDatatypeMismatchError in Phoenix often arises when filtering or comparing columns with incompatible data types (e.g., comparing a string to a number). To fix this, ensure the data types are consistent before applying filters or comparisons, using explicit type conversions if necessary (e.g., converting a string representation of a number to an integer/float using `astype`). Validate your data schema to confirm the correct data types are assigned to each column.
NotImplementedError1 reportThe "NotImplementedError" in arize-phoenix usually means a requested function or method hasn't been defined for a specific class or integration, like a new model type. To fix it, either implement the missing method within the relevant class (e.g., for the Google ADK and Gemini Models, implement the Playground tool logic) or use a supported method if an alternative exists. Sometimes, upgrading to the newest version of the library will resolve this.
Related AI & LLMs Packages
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
Get up and running with OpenAI gpt-oss, DeepSeek-R1, Gemma 3 and other models.
🦜🔗 The platform for reliable agents.
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
LLM inference in C/C++
GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.
Subscribe to Updates
Get notified when new versions are released