← Search

Neon AI Gateway
Database 🧪 Install ran in our test sandbox · Nov 18, 2026 📌 Pinned version: commit 4da86b0
Install test only, on the date shown. This does not guarantee the code is safe.
⚙ Needs: a Neon account on a paid plan (Launch or Scale)…

Neon AI Gateway

One API and one credential for frontier and open LLMs, built into your Neon branch — paid plan, limited regions

At a glance
What
One API and one credential for frontier and open LLMs, built into your Neon branch — paid plan, limited regions
Cost
Free
Needs
a Neon account on a paid plan (Launch or Scale), in one of the supported regions (aws-us-east-2, aws-us-east-1, aws-eu-central-1, aws-ap-southeast-1), with prepaid AI Gateway credits. The gateway is blocked on a Free plan. Image generation additionally needs an OpenAI catalog model for the text call, since the image tool runs on the Responses route, and a supported image model set explicitly (gpt-image-2 by default, or gpt-image-1.5 or gpt-image-1-mini).
Install
Copy the installer prompt below into your Muse — your agent does the rest.

Version:

Quick answer

Neon AI Gateway is a free Muse skill about the Neon AI Gateway, not Postgres, despite its skill ID. It gives one API and one Neon credential for frontier and open-source LLMs from Anthropic, OpenAI, Google, Meta and more, with OpenAI compatibility and model switching by changing one field. Note that a paid Neon plan is required.

⌁

Install

Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 18, 2026📌 Pinned version: commit 4da86b0

Step 0, before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites: a Neon account on a paid plan (Launch or Scale), in one of the supported regions (aws-us-east-2, aws-us-east-1, aws-eu-central-1, aws-ap-southeast-1), with prepaid AI Gateway credits. The gateway is blocked on a Free plan. Image generation additionally needs an OpenAI catalog model for the text call, since the image tool runs on the Responses route, and a supported image model set explicitly (gpt-image-2 by default, or gpt-image-1.5 or gpt-image-1-mini). Install "Neon AI Gateway" for me. One API and one Neon credential for frontier and open-source LLMs, OpenAI-compatible, branch-scoped, plus embeddings and, new in this version, image generation through the OpenAI Responses image tool. Repository: https://raw.githubusercontent.com/neondatabase/agent-skills/4da86b02402ee8d00b3d365a1fccb62a812ec08e/plugins/neon-postgres/skills/neon-ai-gateway/SKILL.md Pinned version: install the code at commit 4da86b02402ee8d00b3d365a1fccb62a812ec08e exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. Updated October 26, 2026: Skill Harbor regenerated this install prompt from the current upstream file. The headline change is image generation. The gateway now documents how to generate images, and the previous prompt did not cover it at all. Images come from the OpenAI Responses image_generation built-in tool on an OpenAI model, on the /openai/v1/responses route. The gateway has no image model endpoint, so the AI SDK generateImage() call throws NoSuchModelError. Only three image models are supported, gpt-image-2, gpt-image-1.5 and gpt-image-1-mini, and the model must be set explicitly on every tool definition, with gpt-image-2 as the default unless you ask for another. Every image request must stream: a non-streaming response whose image is over 655,360 bytes fails with a 400 error, and most images are larger than that, so use streamText, never generateText, when the image tool is attached, or pass stream true with the plain OpenAI SDK. On a Neon Function that streams text and generates images, pick an OpenAI catalog model, since the image tool runs on the Responses route. Nothing else in the gateway changes in this version: the same paid-plan requirement, the same four regions, the same prepaid credits, the same two injected environment variables and no others, and the same branch-scoped credential. The previous prompt also said to copy SKILL.md and its helper files; the skill folder in fact contains SKILL.md only, so this prompt corrects that to SKILL.md alone, and the parent neon skill is a separate skill, signaled here, not installed silently. 1. Fetch the SKILL.md file for the neondatabase-agent-skills-neon-postgres skill from the repository into a temporary folder, then compare SKILL.md with the fingerprint from Skill Harbor's review of 2026-10-26: - SKILL.md: SHA-256 23db230f6c722c9d63dbb6367d40f425426d03bd2b1c38f0978c24b3100864e8 If it does not match, it changed since the review: do not install it yet, tell me what you found and let me decide. 2. Safety check: review the SKILL.md for anything suspicious (unexpected network calls, shell commands, credential harvesting). This skill file should contain zero secrets in code, credentials only via the injected environment variables NEON_AI_GATEWAY_BASE_URL and NEON_AI_GATEWAY_TOKEN, and allowed hosts declared in the file. Verify that holds here; STOP on any red flag and tell me. Tell me in plain words what the skill does when used, which services it connects to, whether it reads or sends any personal information or keys, and anything that seems unusual. Confirm these points from the file: - This is a guidance skill file only: the skill folder contains SKILL.md alone, no helper scripts. It does not install code on my machine by itself; it tells the agent how to call the gateway from an app or a Neon Function, with the Vercel AI SDK, Mastra, or the plain SDKs. - Paid plan and regions first: the gateway is available only on paid Neon plans (Launch and Scale), only in aws-us-east-2, aws-us-east-1, aws-eu-central-1 and aws-ap-southeast-1, and it draws down prepaid AI Gateway credits. A Free plan blocks provisioning; locked models are marked enabled false in GET /v1/models, and the absence of a CLI warning does not mean every model is enabled. - Credentials: Neon injects only NEON_AI_GATEWAY_BASE_URL (the bare branch host, scheme plus host, no path) and NEON_AI_GATEWAY_TOKEN (a Neon credential starting nt_live_). It does not set OPENAI_API_KEY or OPENAI_BASE_URL. Never paste a token value in chat; credentials go through the secure vault or environment variables. - Image generation, the new part: images come from the OpenAI Responses image_generation tool on an OpenAI model, on /openai/v1/responses. The gateway has no image model endpoint, so generateImage() throws NoSuchModelError. Supported image models are gpt-image-2, gpt-image-1.5 and gpt-image-1-mini, set explicitly on every tool definition, gpt-image-2 by default. Always stream: use streamText, never generateText, when the image tool is attached, or stream true with the plain OpenAI SDK, because a non-streaming image over 655,360 bytes fails with a 400 error. - Parent skill: the file says to use the parent neon skill first for the Neon overview and setup. If that parent skill is not installed, tell me and let me decide; do not install another skill silently. 3. Install it as a skill: copy SKILL.md into the agent's skills directory, in a folder named "neondatabase-agent-skills-neon-postgres". 4. Verify with no network calls: frontmatter valid, file in place. 5. Report what was installed, where, and what I still need to do myself (for example: upgrading my Neon plan, creating my project in a supported region, adding prepaid credits, and, for images, picking an OpenAI catalog model and an explicit image model). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it is in the prerequisites above. Rules: do not touch anything outside the temp folder and the install target. Never ask me to paste secrets in chat. Credentials go through the secure vault or environment variables. If anything looks off, stop and ask me.

🔒 Security, provenance and what this touches
🧭 What this touches
Based on the install prompt we scanned, this skill does not run commands, make third-party network calls, install packages, write files, or ask for API keys.
Install prompt analysis · 2026-10-13
🔒 Security and provenance
📌 Pinned version: commit 4da86b0 · 2026-10-13
🧪 Install ran in our test sandbox · Nov 18, 2026
👁 Observed during the install test · Nov 18, 2026 : registry.npmjs.org
🔎 Last checked · Nov 19, 2026 : no change detected in the files linked by this install prompt
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Wanna be extra careful?

The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?

How to check a skill before installing →

✓

Verification certificate

Dated facts about what was checked for this listing, and how. This is not a safety certification.

Pinned version
4da86b0 · 2026-10-13
Sandbox install test
Install ran in our test sandbox · Nov 18, 2026 · Observed during the install test: registry.npmjs.org
Creator identity
Not verified
Fingerprint watch
Last checked · Nov 19, 2026 : no change detected in the files linked by this install prompt
Install prompt analysis
2026-10-13 : No commands, network calls, packages, file writes or API keys detected.

Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify

?

Questions

How do I install a skill?

Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.

Where does my money go?

Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.

What does the ✓ next to a creator’s name mean?

It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.