Owner
pydantic
23 indexed repositories · View on GitHub
-
pydantic
Data validation using Python type hints
Python · 29481 stars
-
pydantic-ai
How Python does AI. Agents, realtime voice, image generation, embeddings. Every model, every interface, typed end to end.
Python · 21172 stars
-
monty
A secure Python sandbox, written in Rust, for code written by AI.
Rust · 8900 stars
-
logfire
AI observability platform for production LLM and agent systems.
Python · 4549 stars
-
httpx2
A next generation HTTP client for Python. 🦋
Python · 1648 stars
-
pydantic-settings
Settings management using pydantic
Python · 1498 stars
-
pydantic-ai-harness
Batteries for your Pydantic AI agent.
Python · 954 stars
-
jiter
Fast iterable JSON parser.
Rust · 573 stars
-
genai-prices
Calculate prices for calling LLM inference APIs.
Python · 412 stars
-
pydantic-extra-types
Extra Pydantic types.
Python · 341 stars
-
speedate
Fast and simple datetime, date, time and duration parsing for rust.
Rust · 255 stars
-
pydantic.run
Python browser sandbox.
TypeScript · 190 stars
-
aioaws
Asyncio compatible SDK for aws services.
Python · 181 stars
-
pytest-examples
Pytest plugin for testing examples in docstrings and markdown files.
Python · 181 stars
-
Python · 153 stars
-
Rust · 78 stars
-
typing-inspection
Runtime typing introspection tools
Python · 73 stars
-
logfire-js
JavaScript and TypeScript SDK for Logfire
TypeScript · 73 stars
-
ai-chat-ui
Web UI for Pydantic AI agents, built over the Vercel AI protocol
TypeScript · 72 stars
-
logfire-helm-chart
Helm Chart for running Logfire 🪵️🔥
Go Template · 60 stars
- 3 more repositories live on View on GitHub
Issues somebody may already be working on are listed last. List everything by date
-
bug pydanty:is-working
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
pydantic/pydantic-ai#13444 ·
Maintainers usually reply within 1 day
-
Unspecified settled-batch ordering changes sandbox errors and continuationsMay be free again A pull request for this issue was closed without being merged. Openpydanty:is-working
Difficulty 3/5 Half a day Newbie friendliness 15/100
pydantic/pydantic-ai#13442 ·
Maintainers usually reply within 1 day
-
Difficulty 5/5 Over a week Newbie friendliness 12/100
pydantic/pydantic-ai#13439 ·
Maintainers usually reply within 1 day
-
flake
Difficulty 4/5 1-2 days Newbie friendliness 22/100
pydantic/pydantic-ai#13424 ·
Maintainers usually reply within 1 day
-
OpenAI-protocol realtime: tool results get no `response.create` after a `response.done` without a readable `response.id`May be free again A pull request for this issue was closed without being merged. Openbug openai realtime
Difficulty 3/5 Half a day Newbie friendliness 50/100
pydantic/pydantic-ai#13410 ·
Maintainers usually reply within 1 day
-
bug realtime
Difficulty 3/5 1-2 days Newbie friendliness 45/100
pydantic/pydantic-ai#13409 ·
Maintainers usually reply within 1 day
-
flake pkg:clai2
Difficulty 3/5 Half a day Newbie friendliness 35/100
pydantic/pydantic-ai#13405 ·
Maintainers usually reply within 1 day
-
Difficulty 4/5 3-5 days Newbie friendliness 22/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
pydantic/pydantic-settings#1150 · 1 comment ·
Maintainers usually reply within 1 day
-
change
Difficulty 4/5 3-5 days Newbie friendliness 45/100
Maintainers usually reply within 1 day
-
Feature Request
Difficulty 4/5 3-5 days Newbie friendliness 56/100
Maintainers usually reply within 1 day
-
feature request
Difficulty 5/5 Over a week Newbie friendliness 25/100
pydantic/pydantic#14272 · 3 comments · 3 reactions ·
Maintainers usually reply within 1 day
-
Aliased fields in `TypedDict` and stdlib-dataclass tool arguments fail durable revalidation on every enginePossibly taken @dsfaccini claimed this today. Opendurable exec pydanty:skip
pydantic/pydantic-ai#13440 · 1 assignee ·
Maintainers usually reply within 1 day
-
`TextOutput` sends the text to a `SkipJsonSchema` parameter named like the `str` parameter's aliasPossibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13438 · 1 assignee ·
Maintainers usually reply within 1 day
-
Track `FunctionSchema.revalidate` in the durable-exec contract so changes to it get `needs-durable-exec-review`Possibly taken @dsfaccini claimed this today. Opendurable exec pydanty:skip
pydantic/pydantic-ai#13437 · 1 assignee ·
Maintainers usually reply within 1 day
-
ReportEvaluator returns Future or Task as an analysis instead of awaiting itPossibly taken @dsfaccini claimed this today. Openbug p:2-high pkg:evals pydanty:triage-completed
pydantic/pydantic-ai#13432 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Render and Azure Durable Task ship nested tool-argument models by Python name, so aliased fields fail revalidation or reach the tool as their defaultPossibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13431 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
A standalone `bool` field's raw probability is thresholded away; `provider_details['probabilities']` has it for every other kind of answerPossibly taken @dsfaccini claimed this today. Openbug needs-maintainer-discussion p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13427 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
`OpenAIDecisionsModel` repeats `Agent(instructions=...)` inside every question after the input, so agent instructions are never part of the cacheable prefixPossibly taken @dsfaccini claimed this today. Openfeature needs-maintainer-discussion p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13428 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
OpenRouter: `provider_details` drops `cost` when OpenRouter reports 0Possibly taken @dsfaccini claimed this today. Openbug needs-maintainer-discussion p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13426 · 2 comments · 1 assignee ·
Maintainers usually reply within 1 day
-
The `Dataset` schema accepts a short form for a multi-parameter evaluator's sole required object-typed parameter, which fails to loadPossibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13421 · 1 assignee ·
Maintainers usually reply within 1 day
-
The `AgentSpec` and `Dataset` schemas accept an empty `{}` capability or evaluator entry, which fails to loadPossibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13420 · 1 assignee ·
Maintainers usually reply within 1 day
-
A mutable tool return stays a live object in history, so mutating it later busts the prompt cache and rewrites the tool resultPossibly taken @its-amann claimed this today. Openbug
Difficulty 4/5 3-5 days Newbie friendliness 35/100
pydantic/pydantic-ai#13412 · 1 comment ·
Maintainers usually reply within 1 day
-
`OpenAIDecisionsModel` raises `UserError` for a request-side `CompactionPart` in an agent runPossibly taken @dsfaccini claimed this today. Openbug p:2-high pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13406 · 2 comments · 1 assignee ·
Maintainers usually reply within 1 day
-
Cross-provider tool-search history: two shapes produce requests that Anthropic and OpenAI reject with HTTP 400Possibly taken @dsfaccini claimed this today. Openbug p:2-high pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13404 · 4 comments · 1 assignee ·
Maintainers usually reply within 1 day
-
clai2: /update breaks run_code in other running sessionsPossibly taken @mpfaffenberger claimed this today. Openclai2 pkg:clai2
Difficulty 5/5 Over a week Newbie friendliness 15/100
pydantic/pydantic-ai#13398 ·
Maintainers usually reply within 1 day
-
S3 deletion sends malformed XML or changes carriage returns in object keysPossibly taken @vinsymoyya claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
-
Rename `pydantic-clai2` to `pydantic-ai-pax`Possibly taken @mpfaffenberger claimed this today. Open
pydantic/pydantic-ai#13393 · 1 assignee ·
Maintainers usually reply within 1 day
-
Sandbox backends differ on whether a background child holding stdout delays `run()`Possibly taken @adtyavrdhn claimed this today. Open
pydantic/pydantic-ai#13391 · 1 assignee ·
Maintainers usually reply within 1 day
-
Report output tool calls skipped by `end_strategy`, so UIs can tell them from the final resultPossibly taken @dsfaccini claimed this today. Openbug needs-maintainer-discussion p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13390 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
OpenAI Responses: an interrupted response sends its `function_call` id after dropping its unsigned reasoning itemPossibly taken @dsfaccini claimed this today. Openbug p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13388 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Empty ThinkingPart replayed as text renders bare `<think>\n\n</think>` in the assistant turnPossibly taken @dsfaccini claimed this today. Openbug p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13387 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Anthropic: a `no_output` retry after a text-only response goes out behind a `'.'` user turn, and Claude answers the `'.'`Possibly taken @dsfaccini claimed this today. Openbug needs-maintainer-discussion p:3-mid pkg:core pydanty:triage-completed
pydantic/pydantic-ai#13386 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Add pydantic-ai-harness in the MongoDB handshake driver metadataPossibly taken @dsfaccini claimed this today. Openfeature p:4-low pkg:harness pydanty:triage-completed
pydantic/pydantic-ai#13384 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Agent workflow provider failure: CI ReviewPossibly taken @dsfaccini claimed this today. Openagentic-workflows pydanty:meta
pydantic/pydantic-ai#13382 · 1 assignee ·
Maintainers usually reply within 1 day
-
The `AgentSpec` JSON schema rejects `NativeTool` capability entries that `Agent.from_spec` loadsPossibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13377 · 1 assignee ·
Maintainers usually reply within 1 day
-
`IPythonCodeMode` timeout can wedge the kernel's shell on Python 3.14 when the interrupt lands inside `print`Possibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13352 · 1 assignee ·
Maintainers usually reply within 1 day
-
`Dataset.to_file()` saves an evaluator's `init=False` dataclass fields, so `Dataset.from_file()` can't load the filePossibly taken @dsfaccini claimed this today. Openbug p:2-high pkg:evals pydanty:triage-completed
pydantic/pydantic-ai#13351 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
A status a retrying transport gave up on loses the provider's quota category: `ModelRateLimitError`, not `ModelQuotaExceededError`Possibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13348 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Run dynamic toolsets (`DynamicCapability` tools) as durable units on AWS Lambda, Azure Durable Task and Absurd, or refuse themPossibly taken @DouweM claimed this today. Openbug durable exec pkg:harness pydanty:skip
pydantic/pydantic-ai#13346 · 1 assignee ·
Maintainers usually reply within 1 day
-
Normalize synthetic capability ids in `ToolDefinition.capability_id` for Prefect cache keysPossibly taken @DouweM claimed this today. Openbug durable exec pkg:core pydanty:skip
pydantic/pydantic-ai#13345 · 1 assignee ·
Maintainers usually reply within 1 day
-
MacAddress rejects addresses with leading/trailing whitespacePossibly taken @Hiyuuna claimed this today. Open
Difficulty 2/5 Under an hour Newbie friendliness 25/100
pydantic/pydantic-extra-types#482 ·
Maintainers usually reply within 1 day
-
`Gate: clean douwebot approval` can't pass a douwebot approval whose `Optional:` finding landed in the review bodyPossibly taken @dsfaccini claimed this today. Openpydanty:skip
pydantic/pydantic-ai#13337 · 1 assignee ·
Maintainers usually reply within 1 day
-
Realtime connection errors with an HTTP status raise an uncategorized `ModelHTTPError`Possibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13336 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
`TextOutput` calls the function with the wrong keyword for an aliased str param next to a hidden param, and for a model-like paramPossibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13331 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
OpenAI image `moderation_blocked` behind a retrying transport that rejects 4xx raises `ModelRequestRejectedError`, not `ContentFilterError`Possibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13330 · 2 comments · 1 assignee ·
Maintainers usually reply within 1 day
-
OpenAI image generation raises an uncategorized `ModelHTTPError` for an OpenAI API errorPossibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13329 · 3 comments · 1 assignee ·
Maintainers usually reply within 1 day
-
`GoogleModel.batch_create` leaves its uploaded input file behind when a retrying transport gives up on the job's 4xxPossibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13328 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
Durable function tool with an aliased parameter fails revalidation with `Field required`Possibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13324 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
`pr-review-gates-events.yml` triggers on `pull_request_review_thread`, which GitHub Actions does not support, so resolving a review thread never wakes `PR Review Gates`Possibly taken @dsfaccini claimed this 1 day ago. Openpydanty:skip
pydantic/pydantic-ai#13322 · 1 assignee ·
Maintainers usually reply within 1 day
Showing the newest 100
This page lists what was indexed most recently. Advanced filter has the whole inventory, narrowed by language, difficulty and how long a task takes.