LOCKMeridian v2.1

OpenAI Assistants API via Meridian

Meridian exposes an OpenAI-compatible endpoint so you can point existing SDKs and toolchains at our infrastructure with zero code changes. Chat completions are fully available as a drop-in replacement. The full Assistants API — threads, runs, tool calls, file search — is currently gated behind an allowlist while we harden state management and concurrency guarantees.

Quick start

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.meridian.sh/v1",
  apiKey: process.env.MERIDIAN_API_KEY,
});

const stream = await client.chat.completions.create({
  model: "meridian-4",
  messages: [{ role: "user", content: "Hello" }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}

Chat Completions

/v1/chat/completions works with streaming, function calling, JSON mode, and vision. Compatible with the official OpenAI Node, Python, and Go SDKs.

Assistants (gated)

Threads, runs, tool outputs, and file search are behind the LOCK gate. Request access via early access.

Authentication

Pass your Meridian API key as a Bearer token in the Authorization header. The OpenAI SDKs handle this automatically when you set apiKey. Keys are scoped per organization; rotate them from the API keys panel.

Need the full Assistants API? assistants@meridian.sh