Pydantic

AI observability can cost more than it should. We ran the numbers: same scale, same workload, four tools. At moderate production load, Logfire is up to 40x less expensive than the alternatives. Spans-based pricing. No proprietary units. AI native and full-stack observability in one platform. See the full comparison: pydantic.dev/articles/ai-ob…

A Pydantic AI agent on your laptop is easy to reason about. Production is less forgiving: a pod dies, a tool times out, a human's offline for approval. New guest post from @htahir1 on Kitaru, a new runtime layer underneath Pydantic AI 👇 pydantic.dev/articles/runti…

We're sponsoring the Model to Market Hackathon by @dawncapital aiengine.build/momq

The ✨️ and most waited version 2 is now in beta.

Pydantic AI version 2.0.0b3 is out! 🎉 github.com/pydantic/pydan…

Pydantic AI version 1.102.0 is out! 🎉 github.com/pydantic/pydan…

Two open source sample apps (Starter Agent and Support Agent) give your team a running head start, with support for @ClaudeAI SDK, @OpenAI SDK, and @Pydantic AI out of the box. Plus, a new Agent Quickstart Guide and overhauled docs make it easier than ever to build, ship, and scale AI agents right from Slack's platform.

Here's a video where I deploy Claude Sonnet from Microsoft Foundry, then build an agent with Microsoft agent-framework, and connect it to a MCP server to order a cupcake (chocolate, of course): youtube.com/watch?v=LYwV8I… (sorry voice is rough, I lost it at PyCon)

"Monty runs cross-platform (no hypervisor or WASM backend), validates LLM-generated code against tool signatures with ty before any host tool fires, and supports Monty-native ResourceLimits for CPU / memory / output caps."

Pydantic AI version 2.0.0b2 is out! 🎉 github.com/pydantic/pydan…

Pydantic AI version 1.101.0 is out! 🎉 github.com/pydantic/pydan…

@pydantic Agents: observable AI pipelines with Pydantic AI + Logfire. Full cost tracking baked in. render.com/templates/pyda…

Try it today: github.com/pydantic/pydan…

- 𝗖𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀 𝗮𝘀 𝘁𝗵𝗲 𝗰𝗼𝗿𝗲 𝘂𝗻𝗶𝘁. One capability bundles an agent's tools, lifecycle hooks, instructions, and model settings, so a whole extension (memory, guardrail, coding toolkit) plugs into every layer at once. - 𝗔 𝗹𝗲𝗮𝗻 𝗰𝗼𝗿𝗲 𝘄𝗶𝘁𝗵 𝗮𝗻 𝗲𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺. Some capabilities ship with Pydantic AI, more from the 1st party Pydantic AI Harness, and others are 3rd party or your own. - 𝗖𝗼𝗻𝗳𝗶𝗴𝘂𝗿𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗼𝗻𝗲 𝗽𝗹𝗮𝗰𝗲. Settings that used to be scattered across `Agent` arguments now live on the primitive. - 𝗠𝗼𝗿𝗲 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆. Updated instrumentation with aggregated token-usage attributes, plus message capture from interrupted runs for easier debugging.

𝗣𝘆𝗱𝗮𝗻𝘁𝗶𝗰 𝗔𝗜 𝘃𝟮 𝗯𝗲𝘁𝗮 𝗶𝘀 𝗵𝗲𝗿𝗲. v2 is built harness-first. Much of what you configure on an agent flows through one primitive, the capability. One unit bundles tools, hooks, instructions, and model settings, so a whole extension plugs into every layer. What's new 👇

Pydantic AI version 2.0.0b1 is out! 🎉 github.com/pydantic/pydan…

Pydantic AI version 1.100.0 is out! 🎉 github.com/pydantic/pydan…

Gregor Stewart is joining Pydantic as an advisor. He's Chief AI Officer at @SentinelOne and has spent his career at the intersection of AI and production systems. We're glad to have him onboard. pydantic.dev/articles/grego…

Pydantic AI version 1.99.0 is out! 🎉 github.com/pydantic/pydan…

We're co-hosting an evening of agent engineering discussions with @GlyphicAI in London on May 20th. @Samuelcolvin and @devang314 will be talking about agents in production. Small room, technical crowd. RSVP luma.com/qqsw0224

🎉 Pydantic (@pydantic) is sponsoring PyCon Italia 2026 as a Startup Sponsor! Welcome on board! 🐍 🤓 Learn more: 🔗 pydantic.dev Tickets 👉 2026.pycon.it #PyConIT #PyConIT2026 #Python