Neon AI Gateway Skill for Claude Code: What It Does and How to Install It

Makes your assistant call LLMs and embedding models through the Neon AI Gateway with one Neon credential.

Installs: 32.4K; on skills.sh
Stars: 108; on GitHub
Updated: Oct 2026; last repo commit

SKILL.md enables the gateway with aiGateway: true in neon.ts, deploys it, and wires the injected token and base URL into the Vercel AI SDK, Mastra or the plain OpenAI, Anthropic and Gemini SDKs. It also covers images, embeddings and listing models.

It loads when you ask to call an LLM, add AI to an app or route between model providers. It needs the Neon CLI, a paid Neon plan with prepaid credits and a project in one of four AWS regions. It points to the parent neon skill first.

Good for

Developers adding chat or agents to an app that already uses Neon
Teams avoiding separate provider accounts and API keys
People using the Vercel AI SDK or Mastra

Install neon-ai-gateway

Claude Code

Copy this prompt.

Install the neon-ai-gateway skill in Claude Code. Run this command:

npx skills add neondatabase/agent-skills --skill neon --skill neon-ai-gateway

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js, A Neon account.

Install the parent neon skill with this one: SKILL.md starts with 'Use the parent neon skill', and the commands here use the Neon CLI. The parent neon skill gives the CLI install (npm i -g neon, plus bun, pnpm or brew variants) and sign-in (neon login). The plugin route installs the neon-postgres plugin: 9 skill folders and the Neon MCP server (https://mcp.neon.tech/mcp), which asks you to sign in with Neon over OAuth on first use. The folder routes copy skills only. There is no Claude app upload route because the skill drives the Neon CLI.

Built from the package data we hold. Not yet tested by us.

What's in the skill

When it loads

One API and one credential for frontier and open-source LLMs and embedding models, built into your Neon branch and powered by Databricks. Use when a user wants to call an LLM, generate embeddings, add AI/chat/an agent to their app, route between model providers (OpenAI, Anthropic, Google/Gemini, Meta, Alibaba, and more), or avoid juggling separate provider API keys and accounts — especially when they already use Neon and want AI requests to branch with their project. Works with the OpenAI SDK, Anthropic SDK, google-genai, the Vercel AI SDK, and Mastra by changing only the base URL. Triggers include "call an LLM", "add AI to my app", "chat completion", "model routing", "LLM proxy/gateway", "one API for all models", "use Claude/GPT/Gemini", "AI SDK", "Mastra agent", "Neon AI Gateway", "generate embeddings", "generate images", and "log/rate-limit AI calls".

Sections in SKILL.md

When to Use
What It Does
Availability
Enabling the gateway: plan and model access
Setup
Neon Infrastructure as Code (neon.ts)
Environment Variables
Build Agents with the Vercel AI SDK (Recommended)
Generate Images
Generate Embeddings with the AI SDK
Build Agents with Mastra (Recommended)
Use with Plain SDKs (Lower-Level)
Model Identifiers
List Available Models at Runtime (/v1/models)
Neon Documentation
Further Reading

Key facts

Author: Neon (neondatabase/agent-skills, vendor's own repo)
Repo: neondatabase/agent-skills
License: Apache-2.0 (LICENSE and plugin.json agree)
Installs on skills.sh: 32,351 (Nov 2, 2026)
GitHub stars: 108 for the whole neondatabase/agent-skills repo (Nov 2, 2026)
Last repo commit: Oct 30, 2026
Audits on skills.sh: Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass
What it includes: 1 file: SKILL.md (counted from the repo tree)
Needs: The parent neon skill, the Neon CLI and a Neon account (the parent skill gives npm i -g neon)
Part of plugin: neon-postgres (9 skill folders and the Neon MCP server)

What it includes

Four AWS regions; paid Neon plans (Launch and Scale) with prepaid AI Gateway credits.
SKILL.md says neon config apply, deploy and checkout refuse to enable the gateway on a Free plan.
Read · Availability and plan check
Enables the gateway on a branch with neon deploy.
Each branch gets its own gateway host.
Try: “Add AI chat to my app with the Neon AI Gateway.”
Runs code · aiGateway: true in neon.ts
NEON_AI_GATEWAY_TOKEN and NEON_AI_GATEWAY_BASE_URL, written by neon env pull.
Runs code · Environment variables
Builds agents with the AI SDK using the @neon/ai-sdk-provider.
Write · Vercel AI SDK agents
Image generation and embeddings through the gateway with the AI SDK.
Write · Generate images and embeddings
Builds agents with Mastra using neon/<model> identifiers.
Write · Mastra agents
OpenAI, Anthropic and Gemini SDKs with the gateway base URL.
Routes are /v1, /openai/v1, /anthropic and /gemini.
Write · Plain SDKs
Reads GET /v1/models, which marks locked models enabled: false.
Network · List models at runtime

From the official docs

Permissions

Medium risk

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

It spends your AI credits
SKILL.md says the gateway is for paid Neon plans (Launch and Scale) and draws down prepaid AI Gateway credits. Streaming agents and image generation can use credits quickly. github.com
A bearer token lands in your env file
neon env pull writes NEON_AI_GATEWAY_TOKEN to .env. Keep .env out of git, and do not ship the token to browser code. github.com
Your prompts go through the gateway
Model calls are sent to the branch gateway host, which routes them to the model provider you choose. Do not send data you cannot share with that provider. github.com

Technical details

Turning on read-only	No read-only mode is documented for this skill.
Talks to	Neon AI Gateway host (https://<branch-id>-api.ai.<region>.aws.neon.tech). Neon API through the CLI
Sign-in and keys	Uses your Neon CLI sign-in or API key. The gateway token is written to .env by neon env pull.
Scopes	Runs the Neon CLI as your Neon account Writes a gateway bearer token into your .env Model calls draw down prepaid AI Gateway credits on a paid Neon plan Installs AI SDK packages for the code it writes

Questions

What does the neon-ai-gateway skill do?

It helps your assistant send LLM and embedding calls through the Neon AI Gateway, with one Neon credential for models from several providers. It enables the gateway in neon.ts and wires it into the AI SDK, Mastra or the plain SDKs.

How do I install it?

Use the install box at the top of this page. It installs the parent neon skill with this one, because SKILL.md starts by pointing to the neon skill. You also need the Neon CLI and a Neon account.

Is it the official Neon skill?

Yes. It is in Neon's own agent-skills repo, and the repo's marketplace file calls its plugins official Neon plugins for Claude Code.

Does it work in the Claude app?

The install box has no Claude app upload route, because the skill drives the Neon CLI on your machine.

Does it cost money?

The gateway needs a paid Neon plan (Launch or Scale) and draws down prepaid AI Gateway credits, per SKILL.md. It refuses to enable the gateway on a Free plan.

Do I still need OpenAI or Anthropic keys?

SKILL.md says Neon injects only NEON_AI_GATEWAY_TOKEN and NEON_AI_GATEWAY_BASE_URL and does not set OPENAI_API_KEY. You use that one Neon credential instead of separate provider keys.

Which SDKs work?

SKILL.md lists the OpenAI SDK, Anthropic SDK, google-genai, the Vercel AI SDK and Mastra, by changing only the base URL.

Is it safe?

It spends credits, writes a token to .env and sends model calls through the gateway. Our advice: keep the token on the server and watch your credit balance.

Information

updated Nov 2, 2026

Repository: neondatabase/agent-skills
Folder: skills/neon-ai-gateway

neondatabase/agent-skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 6

neondatabase/agent-skills README · github.com
neon marketplace.json · github.com
neon plugin.json · github.com
neon mcp.json · github.com
neon SKILL.md (parent skill) · github.com
neon-ai-gateway SKILL.md · github.com