Skip to main content

​
Install

npm install @maximem/synap-vercel-adk

​
What’s included

ExportPurpose
createSynapAsync factory that initializes the Synap provider
SynapProviderProvider class with wrap and listen methods

​
Quick start

import { generateText } from "ai";
import { anthropic } from "@ai-sdk/anthropic";
import { createSynap } from "@maximem/synap-vercel-adk";

const synap = await createSynap({
  apiKey: process.env.SYNAP_API_KEY!,
  instanceId: process.env.SYNAP_INSTANCE_ID!,
});

const model = synap.wrap(anthropic("claude-sonnet-4-6"), {
  userId: "alice",
  customerId: "acme",   // optional
});

const { text } = await generateText({
  model,
  messages: [{ role: "user", content: "What do you remember about my account?" }],
});
synap.wrap returns a standard Vercel AI SDK LanguageModel — pass it anywhere you’d use a plain model.

​
How it works

On every generateText / streamText / generateObject call:
  1. Before — fetches the user’s Synap context and injects it as a system message
  2. Generates — proxies the request to the wrapped model unchanged
  3. After — ingests the completed user + assistant turn into Synap asynchronously
your code → synap.wrap(model) → [fetch context] → wrapped model → [ingest turn] → your code

​
Works with any model

Wrap any Vercel AI SDK-compatible model: 
import { openai } from "@ai-sdk/openai";
import { google } from "@ai-sdk/google";
import { anthropic } from "@ai-sdk/anthropic";

const gptWithMemory    = synap.wrap(openai("gpt-4o"),                 { userId: "alice" });
const geminiWithMemory = synap.wrap(google("gemini-2.0-flash"),        { userId: "alice" });
const claudeWithMemory = synap.wrap(anthropic("claude-sonnet-4-6"),    { userId: "alice" });

​
Streaming

Works with streamText and streamObject without any changes: 
const { textStream } = await streamText({
  model: synap.wrap(openai("gpt-4o"), { userId: "alice" }),
  messages: [{ role: "user", content: "Summarize my recent priorities." }],
});

for await (const chunk of textStream) {
  process.stdout.write(chunk);
}

​
Per-request scoping

Override the user scope per call by wrapping per-request: 
async function handleChat(userId: string, message: string) {
  const model = synap.wrap(openai("gpt-4o"), { userId });
  const { text } = await generateText({
    model,
    messages: [{ role: "user", content: message }],
  });
  return text;
}

​
Anticipation stream (advanced)

synap.listen() opens a gRPC stream that pre-fetches context speculatively before the user’s request arrives, reducing perceived latency: 
const stop = synap.listen({ userId: "alice" });

// Later, when done:
stop();
Use listen in long-lived server processes where you can predict who will send the next message.

​
Next steps

Mastra

SynapMemory and tools for the Mastra ADK.

Claude Agent SDK

Hooks and MCP server for the Claude Agent SDK.