NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #5026 most downloaded on PyPI
A helper library to interact with Arize AI APIs
Last release 5 days ago
29 Sep 2026
Ships on a steady schedule
a new release about every 2 weeks
Nearly every release is documented
notes for 59 of the last 60 stable releases
10 versions withdrawn
withdrawn after publishing
7 years old
367 releases · first in 2020
One column per quarter.
datasets: adding complex filter queries to list dataset examples
datasets & experiments: accept Parquet/Arrow file paths in create() , streamed so large files don't need to fit in memory
integrations: OAuth 2.0 client credentials for custom endpoints
integrations: fireworks support in the hand-written integration subclients
auth: derive REST user permissions from account roles
annotations: Add Python SDK support for session annotations
spans: add column projection to list spans
traces-python: add client.traces.list for /v2/traces ( #82771 ) ( 885d6c2 ), closes #82765
separate task run configuration request schemas
evaluations: pin eval tasks to a specific evaluator version
classification_choices and default data_granularity on the REST eval endpoint; regenerate SDKs to match (#79714) (9bd254d)TaskRun API responses, so new server fields no longer break existing clients (#82010) (af6f709)GitOrigin-RevId: 9c4efb7d79f44dfe3841d3d0de27fbc14378659a
GitOrigin-RevId: 9c4efb7d79f44dfe3841d3d0de27fbc14378659a
GitOrigin-RevId: b1bc43b776ac74ffa90ccdceedea7c6965860373
GitOrigin-RevId: b1bc43b776ac74ffa90ccdceedea7c6965860373
GitOrigin-RevId: 4d7f543c150d6570c823ab7521667a8a1987d15b
GitOrigin-RevId: 4d7f543c150d6570c823ab7521667a8a1987d15b
GitOrigin-RevId: 07595525a6464844ef9a9295b8fc0c089f4572f5
GitOrigin-RevId: 07595525a6464844ef9a9295b8fc0c089f4572f5
GitOrigin-RevId: b0dde7af5cebf84fe99cc1ecc95a706ccd56ca1f
GitOrigin-RevId: b0dde7af5cebf84fe99cc1ecc95a706ccd56ca1f
GitOrigin-RevId: 51c3be27fff1e223e7ddd142b3ddeaaa924da5bb
GitOrigin-RevId: 51c3be27fff1e223e7ddd142b3ddeaaa924da5bb
Minor release. The v2 REST API standardization (#78907) is technically breaking, but only affects endpoints/methods in
alphaorbeta— all gated behind the pre-release opt-in and documented with a warning. No stable surface changes.
SCREAMING_SNAKE_CASE — AI integrations, prompts, evaluators, tasks, orgs, spaces, users, API keys, and roles (#78718, #78720, #78721, #78722) (b17b78d)operationId naming (*ListResponse → List*Response), incl. annotation-config & evaluator-version request schemas (#79270, #79433) (b17b78d)set-prompt-version-labels now returns the full PromptVersion (#79278) (b17b78d)404 for list endpoints when the scoped resource is missing or inaccessible (#79279) (b17b78d)resource-restrictions: add list functionality
delete_examples (#77131) (a713b5a)audit-logs: GET /v2/audit-logs — paginated audit log retrieval for enterprise accounts
Spans For delete, structured 200 with completed/deleted/not_deleted, 503 for errors
datasets: add cursor pagination for List Dataset Examples
annotations: promote annotation /annotate endpoints to beta + cap queue annotate array
api-keys: removed delete method
sdk-config: add default_headers support
experiments: support for appending experiment runs
add missing Bulk delete response object
users: give bulk-delete results separate UserID and Email fields
prompts: add get version support
promote annotation-queues, organizations, prompts, roles, spaces to BETA
projects: support for update project
resolve type-check errors in users, api-keys, evaluators, annotation-configs, and tasks
experiments: populate experiment span metadata with dataset and experiment keys
extract inline response schemas to named schemas
prompts: add SDK wrapper types with clean __str__ for CLI display
prompts: add missing fields breaking prompts.get() (#72583) (6954a17), closes #72582
types: extract inline enums for annotation_queues
api-keys: ArizeClient.api_keys.create_service_key() and full API keys docs
ArizeClient.api_keys.create_service_key() and full API keys docs (#71842) (6090c4e)ArizeClient.users.bulk_delete() bulk convenience method (#71843) (e87668a)ArizeClient.users.find(email=...) for exact email lookup (#71841) (d103eb7)annoation_configs: add inline enums
Add structured annotations to dataset examples and experiment runs responses
users: Add users management support
api-keys: extend GET /v2/api-keys with space_id and user_id filters
add missing pre-defined metadata headers
sdk: Respect request_verify for REST API commands
annotation-queues: Update how "unset" behaviour works for annotation queues.
arrow: normalize legacy 'result' column to 'output' across schemas
extend evaluators API for code evaluator configs
experiments: add force_http fallback
types: add missing public type re-exports to types.py files
organizations: add organizations.delete
annotations: add annotate_examples and annotate_runs methods
add DatasetWithExampleIds response for dataset examples endpoints
add space name/ID resolution to api_keys.create()
organizations: add OrganizationsClient with list/get/create/update support
roles: support name or ID in RolesClient.update()
spaces: add delete space support
resource-restrictions-and-role-bindings: add ResourceRestrictionsClient and RoleBindingsClient subclients
roles: support name_or_id in RolesClient.get() and .delete()
Your coding agent can read these notes before it upgrades. Set up the MCP server →