We're taking part in Stripe Projects. You can now provision Pydantic Logfire straight from the command line while you set up a project. No signup form, no email verification, no tabbing between dashboards:
stripe projects init
stripe projects add pydantic/logfire
A Logfire project is provisioned in your own account, and your project credentials are written straight to your environment:
$ stripe projects add pydantic/logfire
○ Provisioning pydantic/logfire...
├─ ✓ Resource requested
├─ ✓ Resource provisioned
├─ ✓ Credentials synced
└─ ✓ Project updated
● pydantic/logfire ready
│ ✓ Added "pydantic-logfire" to active environment "default" (.env)
│ ✓ 5 credentials created for Pydantic:
│ PYDANTIC_BASE_URL=http••••••••
│ PYDANTIC_ORGANIZATION_ID=9aab••••••••
│ PYDANTIC_ORGANIZATION_NAME=••••••••
│ PYDANTIC_PROJECT_ID=2a5b••••••••
│ PYDANTIC_PROJECT_NAME=••••••••
Stripe Projects is a CLI workflow for discovering, provisioning, and managing the services your app depends on, straight from the terminal. The provisioning protocol is one we co-designed with Stripe: it provisions a real service and hands back real credentials, deterministically, without a human tabbing between dashboards to make it happen. The account stays yours, with the access and controls you already expect.
Pick a plan without leaving the terminal
Billing runs through your Stripe project. When you add Logfire, the CLI asks which plan you want, and you pay through Stripe with no separate billing setup on the Logfire side:
- No plan: standalone pricing. Provision the service without a plan and start sending data on the free tier.
- Team, $49 per month: for startups and small teams shipping to prod. 5 seats included (up to 12, $25 per extra seat), 5 projects, 10M records a month included, 30-day retention.
- Growth, $249 per month: for scaling teams. Enough seats, guests, and projects for your AI application, up to 90 days retention, priority support, self-service data deletion, and a HIPAA BAA.
Observability should be there from the first run
When you ship AI features, you can't fully predict how they behave in production. An agent takes a path you didn't expect, an LLM call comes back wrong, a tool fires with the wrong input. Pydantic Logfire shows what your AI features and agents actually do once real traffic hits, which is how you catch problems no test would've caught. Observability is how you find out whether the thing works.
The catch is that setting it up was mostly manual, and it was the one step a coding agent couldn't do for you. It can write the application, but it can't click through a dashboard flow or read a token off a screenshot. It needs deterministic steps and a real credential. Miss either, and observability becomes the step that gets skipped.
Provisioning Logfire through Stripe Projects closes that gap, for people and agents alike. stripe projects init scaffolds skill files into your repo that teach Claude Code, Cursor, or any coding agent how to use the CLI.
├── .agents
│ └── skills
├── .claude
│ └── skills
├── .cursor
│ └── rules
├── .cursorignore
├── .env
├── .gitignore
...
├── AGENTS.md
└── CLAUDE.md
Your agent runs the same two commands and your project is provisioned, with its credentials in the environment. Here is how you could prompt it:
You: "Add logfire observability to this project"
Agent: [runs stripe projects add pydantic/logfire]
Agent: "Done. Logfire is provisioned and your credentials are in the environment.
Want me to instrument the app with the Logfire SDK?"
The code your agent writes is exactly the code you want traces for. Now the agent that writes it can also wire up the tool that watches it.
From there, instrument as usual
import logfire
logfire.configure()
logfire.info('hello from a project that provisioned its own observability')
Logfire is built on OpenTelemetry, so the JavaScript and Rust SDKs work similarly, as well as any other standard OTel SDK.
Getting started
Provisioning Logfire through Stripe Projects is currently US-only. It's ideal for teams building AI features and agents, and for anyone who wants a shorter path from a new repo to a running project.
-
Initialize your project and add Logfire, picking a plan when prompted:
stripe projects init
stripe projects add pydantic/logfire
- Instrument your app with the Logfire SDK and start reading traces.
Whether a developer or a coding agent stands up the project, observability goes up with it in the same command. Your app is observable from the first request instead of when someone remembers to add Logfire.
To try it, install the Stripe CLI and run stripe projects init.
