From the team behind Pydantic, Logfire is a new type of observability platform built on the same belief as our open source library — that the most powerful tools can be easy to use.
External link for Pydantic
California, US
Pydantic reposted this
This is EXACTLY why Pydantic AI Gateway DOESN'T try to convert LLM requests to a single API - if you want to make full use of a model, you'll need direct access to the full real API as soon as new features land. Instead we do the boring stuff like auth, budget limits, monitoring, and otherwise get the f**k out of your way.
Pydantic Logfire shining at the Manchester AI community. Using Logfire in production? Reach out and tell us how.
The third Practically AI meetup was another successful night for Manchester's AI community! 🚀 I had the privilege to speak about building agentic AI applications with the OpenAI Agents SDK, including tracing and debugging with Pydantic Logfire - the questions afterwards showed just how many people are tackling these challenges in production. Lewis Murphy showed us how BNY is putting agentic AI in everyone's hands with Eliza - non-technical staff building their own agents. Proper democratisation. A huge thank you to: • Iulii Cebotari, Edwin Luther and BNY for hosting us and providing such a fantastic venue at short notice • Phil Whittaker for bringing Practically AI to life and giving me the opportunity to contribute • My colleagues Evgeny Lukiyanov and Kveto Tulipan at AZoNetwork for their continuous support • Organizers Philip Simpson, Adam Robinson and Michael Collins (iO Associates) for all the help • Former colleagues and friends who came out to support • Everyone who showed up, asked questions, and stuck around to chat Already looking forward to the next one. If you're interested in future events focused on practical AI applications, join our meetup page to stay in the loop https://lnkd.in/eJsuTu2a Thank you all for making Manchester's AI community thrive! 🙌 #AI #GenAI #AgenticAI #MachineLearning #TechCommunity #Manchester #PracticallyAI #SoftwareEngineering
We got together with Temporal Technologies to show how to bring durable execution to type-safe AI agents. The live coding session was about how to create agents that support API failures, can handle long-running workflows, and coordinate multiple agents without manual checkpoints or fragile retry logics. Thank you to Shy Ruparel for hosting. And to our CTO David M. and dev lead of Pydantic AI Douwe Maan for participating in the webinar. Curious? The recording link is in the comments section. Want to learn more about building durable agents with Temporal, Pydantic Logfire, and Pydantic AI? Checkout the blogpost https://lnkd.in/gp3vxbga #webinar #livecode #pydantic #logfire #ai #temporal #recording #durableagents
Building production AI agents reveals that architectures that work for simple demos become impossibly painful at scale. When long-running workflows fail, you lose all compute, progress, and user trust. The solution is to build systems leveraging durable agents that can recover from failures instead of forcing users to start over. Samuel Colvin gave a talk at @aidotengineer explaining the ins and outs of building applications with Pydantic AI & Logfire with Temporal Durable Agents. https://lnkd.in/e9SeCrgY #pydanticlogfire #durableagents #temporal #pydanticai
Pydantic reposted this
Pydantic team just launched the Pydantic AI Gateway (PAIG) in Beta, and it’s a big step forward for anyone building multi-provider LLM systems. One of the recurring challenges in LLM development is dealing with: - different API formats for each provider - fragmented cost tracking - limited observability - manual failover logic when a model is unavailable PAIG simplifies all of this with one unified interface. • Single API key for all major LLM providers (OpenAI, Anthropic, Google Vertex, Groq, AWS Bedrock) • Real-time cost monitoring + spending limits Daily/weekly/monthly caps at the user, project, or key level. • Built-in observability Full request logging via OpenTelemetry / Pydantic Logfire. • Automatic failover & multi-provider routing If a model fails, PAIG can seamlessly switch to another provider. If you ask me how it differs from other LLM providers, I’d highlight two things: • It integrates seamlessly with the Pydantic and Pydantic-AI ecosystem, giving you structured outputs, validation, agent workflows, and full observability without any additional setup. • It follows a zero-translation approach, sending requests in each provider’s native API format, which means new model features become available immediately, with no schema adjustments or compatibility delays. And of course, huge thanks to the Pydantic team and Samuel Colvin for building an ecosystem that consistently goes beyond expectations. #AI #LLM #Agent #Pydantic
We had a blast at #FastMCP's birthday party last week in New York! Thank you Clay for hosting us in your office space. And thank you to our guests speakers: Rohil H., Cesar Duran, Adam Azzam, and Samuel Colvin. Checkout the highlights of the party in the video below. #mcp #prefect #pydanticlogfire #gtm #clay #meetup #pyAI #pydantic #ramp
Last week, we took over Clay's office for the evening with talks by: Samuel Colvin from Pydantic, Rohil H. from Clay, our own Adam Azzam, and Cesar Duran from Ramp. The presentation on Ramp Research had people standing around long afterward still talking it through. Turns out engineers love it when you tell them what you built and why instead of whatever we normally do at these things. This whole series started in SF a few weeks back and it's been growing. People stay way after, conversations keep going. London is next. December 16th at Google Cloud's London office, Pydantic is hosting. There will be mulled wine and talks from Samuel Colvin, WALT Labs, and SurrealDB. If you're on that side of the world, come! Link in comments 👇
Pydantic reposted this
The Pydantic & Temporal teams are going live at 1pm EST / 6pm GMT *today* to walk through building durable AI agents with Pydantic AI and Temporal Technologies. They will do live coding to show: * create agents that survive API failures * handle long-running workflows * coordinate multiple agents without manual checkpointing or fragile retry logic. Sign up here: https://luma.com/jj50rou0
The Pydantic & Temporal teams are going live at 1pm EST / 6pm GMT *today* to walk through building durable AI agents with Pydantic AI and Temporal Technologies. They will do live coding to show: * create agents that survive API failures * handle long-running workflows * coordinate multiple agents without manual checkpointing or fragile retry logic. Sign up here: https://luma.com/jj50rou0
The first edition of the Py AI meetup was a success and we can prove it! Checkout the highlights of the first ever edition. And if FOMO hits you, join us in NY (https://luma.com/9fpk2sil), London (https://luma.com/2l9bwqto) or keep an eye out for the Py AI conference next year. https://lnkd.in/e8Rp7PmB
We kicked off the PY AI meetup series in SF last week and came out swinging. World-first announcements, groundbreaking demos, and AYCE pizza set the bar high. Join us for the next meetup THIS FRIDAY, November 21st in NYC! Hear from: Samuel Colvin — Pydantic Rohil H. — Clay Jeremiah Lowin — Prefect & FastMCP Cesar Duran — Ramp 6PM at Clay's NYC office. More pizza, talks, and networking with the Python AI community. Link below. See you there!
Pydantic reposted this
Built right.
Pydantic AI Gateway: Open Beta Now Live 🚀 https://lnkd.in/emyhdCd9 Introducing Pydantic AI Gateway (PAIG) — AI model plumbing, the Pydantic way. One key, access to all models. Zero translation, requests flow through in the provider's native format. When OpenAI, Anthropic, or Google ship something new, you use it immediately. What you get: - One key for OpenAI, Anthropic, Google Vertex, Groq, and AWS Bedrock - Cost limits at project, user, and API key levels (daily, weekly, monthly caps) - Built-in observability through Logfire or any OpenTelemetry backend - Automatic failover around provider outages - Open source core (AGPL-3.0) with self-hosting option - Enterprise ready: SSO, granular permissions, flexible deployment Free during open beta. No waitlist. Sign up at https://lnkd.in/ev8kr4Fg Blogpost with announcement: https://lnkd.in/echee3Ja #pydanticaigateway #aigateway #pydantic