Fly 在 2026 年发布的内容

Tag All the Things!

Fly.io released a technical guide showing how to implement tagging for Phoenix applications using Postgres array columns and Ecto. The post details database migrations, schema definitions, validation,…

Struct: A Machine Per Agent, Thousands at a Time

Struct, a Fly.io customer, adopted Fly Machines to run thousands of isolated agent sandboxes for incident response, ensuring hard multi-tenant security and sub-second startup times. Each agent runs in…

Kiloclaw: Hosting Thousands of Claws

KiloClaw, an AI agent platform, chose Fly.io to host thousands of OpenClaw instances due to its always-on machines, strong isolation, and security model. The case study emphasizes Fly.io’s ability to …

Starchild: Vibe Trading Requires Infrastructure That Can Keep Up

Starchild, a trading platform, leverages Fly Machines to run modular AI agents for vibe trading, suspending compute between tasks to reduce idle costs. Fly’s networking and isolation handle multi-agen…

Dat AI Systems: From Concept to Scale in Record Time

DAT AI Systems, a Pakistani startup, built a field operations solution for Lahore Waste Management Company using Fly.io’s infrastructure, Elixir, and serverless components. The system supports over 4,…

How Hatchet Keeps Background Tasks from Ruining Your Day

Hatchet, an open-source background task service, credits Fly.io’s Machines infrastructure for enabling stateful containers, multi-region scaling, and seamless deployments. The lean team avoided DevOps…

How Zapi Makes Building Internal Tools Less Painful

Zapi, a tool for turning PRDs into working prototypes, credits Fly.io’s Machines for instant stateful environments and fly-replay for eliminating routing complexity. The lean team avoids infrastructur…

How Steel.dev Turns AI Agents Into True Browser Ninjas

Steel.dev, an AI-driven browser automation platform, credits Fly.io’s Machines API for enabling rapid, scalable deployment of cloud browsers without Kubernetes complexity. Their orchestrator leverages…

How Turso Made SQLite Harder, Better, Faster, Stronger

Fly.io’s infrastructure enables Turso to offer a production-grade SQLite variant (libSQL) with replication, backups, and vector search, while leveraging Fly.io’s scale-to-zero and global deployment ca…

How Tigris Data Makes Object Storage Screaming Fast, Globally

Tigris Data, an S3-compatible object storage provider, announced its integration with Fly.io’s infrastructure, enabling globally distributed apps to leverage low-latency object storage. Built on Found…

Pricing change detected for Fly

Pricing updated for Fly: - New tier: HIPAA Compliance Package ($99/mo)

Pricing updated for Fly: - New tier: Standard Support ($29/mo) - New tier: Premium Support ($199/mo) - New tier: Enterprise Support ($2500/mo) - Removed: Serious Support tier - Removed: Compliance Mad

Your Agent Speaks MCP. Give It a Computer.

Fly.io introduced MCP support for its Sprites disposable cloud computers, enabling agents to interact with Sprites via a structured protocol. The update includes safety annotations for operations, pro…

How CodeCrafters Helps Devs Upskill Faster

CodeCrafters, a structured learning platform for developers, moved its code evaluation workloads from in-house Firecracker VMs to Fly.io to handle spiky traffic without over-provisioning. Using Fly.io…

How MyMahi Moves Fast Enough to Keep up With Teenagers

MyMahi, a New Zealand-based education platform used in over 100 schools, migrated its infrastructure to Fly.io to handle unpredictable data syncs from large schools. The move reduced DevOps costs and …

Tigris Data introduced a new object storage solution built on FoundationDB, offering S3-compatible APIs with global buckets and real-time performance. Unlike traditional object stores, Tigris caches a

Fly.io introduced Turso, a platform enabling SQLite use in modern production workloads by adding replication, backups, disaster recovery, and native vector search. Turso is built on Fly.io's infrastru

Steel.dev, an AI-driven browser automation platform, adopted Fly.io's Machines API to replace Kubernetes and Firecracker VMs, enabling rapid, scalable deployment of cloud browsers for AI agents. The p

Zapi, a tool for turning product requirement documents into working prototypes, highlights Fly.io’s Fly Machines and fly-replay for instant stateful environments and simplified routing. The case study

Hatchet, an open-source service for high-throughput background tasks, highlights its use of Fly.io’s Machines infrastructure to avoid serverless limitations like rigid timeouts and dropped tasks. The