Neon

Neon支持数据库分支（Database Branching），开发者可以为每个PR创建独立的数据库副本进行测试

Neon是新一代Serverless PostgreSQL服务，其核心创新在于存储与计算分离的架构，支持在无请求时将计算层缩容至零（scale-to-zero）

The default models of multiple Providers have changed: Cortecs, CrossModel, Eden AI, Google Gemini, GitHub Copilot, Google Vertex AI, LLM Gateway, Ofox, OpenCode Go, and Prime Inference now default to Gemini 4 Argon; Fireworks AI, Nebius Token Factory, and Requesty switch to Muse Spark Open; CoreWeave, DigitalOcean, NanoGPT, and Nvidia adopt Nemotron 3.5 Super VL; Aixy and Neon use Claude Haiku 5.5; Agnes AI, Ambient, IteraCompute, Poe, Scaleway, and SenseNova also have their own adjustments.

The database branch addresses data isolation and orchestration problems, and does not improve the model's own forecasting capability.

v0's full flow required 6 manual steps: accept the terms, install Neon, approve SQL twice, generate the auth secret, and publish manually.

HireLens AI's deployment architecture is: frontend on Vercel, backend on Render, PostgreSQL database on Neon, all using free-tier infrastructure.

Render's free web services are suspended after a period without traffic, and the next request triggers a cold start that takes tens of seconds to bring the process back up; Neon's free PostgreSQL likewise suspends its compute node after idling, so the first connection has a wake-up delay.

Lumen OS follows a strict local-first philosophy: all active telemetry data, trade logs, risk metrics and interface state are kept in local memory and synced to browser storage, with data synced asynchronously to a Neon serverless PostgreSQL database, so the UI does not block or freeze even if the network is interrupted or packets are dropped.

Prime Inference's default model is Gemini 4 Argon, Requesty's is Muse Spark Open, Nvidia's and Nebius Token Factory's is Nemotron 3.5 Super VL, and Aixy's and Neon's is Claude Haiku 5.5.

Nikita Shamgunov is a co-founder of Neon and previously also founded MemSQL (later renamed SingleStore).

Neon is a cloud database company providing Serverless Postgres, characterized by separating PostgreSQL's storage from compute and supporting on-demand scaling, branched databases, and extremely fast cold starts.

Single-file configuration has boundaries: as application complexity rises, whether one file can carry the complete backend logic depends on the thickness of the abstraction layer; Neon positions this file as a configuration entry point rather than a container for business logic.

Writing the Neon configuration into neon.ts makes the configuration file itself code, which can go into version control, participate in code review, be checked by the type system, and be deployed along with the application, thereby reducing configuration drift.

Neon's architecture separates the storage layer from the compute layer, with data stored on shared storage in a copy-on-write manner, and compute nodes mounting to a particular data version on demand.

Creating a branch in Neon is essentially just adding a pointer to a particular data state, yielding a fully functional, readable and writable independent database within seconds, with almost no additional storage overhead.

Neon branches are isolated from each other, and a schema migration or data write on one branch does not affect the parent branch.

Neon is a Serverless PostgreSQL service based on a storage-compute separation architecture; its branching feature uses the Copy-on-Write concept and can create isolated environments containing a full dataset within seconds.

In the demo, TabFM runs locally on a DGX Spark, with data stored in a Neon-based Serverless Postgres.

The solution to the data leakage is to have the agent perform training on an isolated branch restored to the exact timestamp prior to the prediction window.

Serverless Postgres products such as Neon and PlanetScale attempt to overcome the database cold-start bottleneck using storage-compute separation and instant branching technology.