User Profiles

Fetch and use automatically maintained user context

User profiles are extremely short summaries of context about an entity (Usually a user, but can be anything) which includes both the static facts about them, as well as a few recent episodes.

You can think of these as a dynamic compaction that's done by supermemory in real-time.

This profile should be injected into the agent context for truly personalized experiences. To read more, visit User profiles - Concept

Get a user's profile — their static facts and dynamic context — with a single API call.

TIP

Profiles are built automatically as you ingest content. No setup required.

Quick Start

TypeScript: ```typescript import Supermemory from 'supermemory';

const client = new Supermemory();

const { profile } = await client.profile({
  containerTag: "user_123"
});

console.log(profile.static);   // Long-term facts
console.log(profile.dynamic);  // Recent context
```

Python: ```python from supermemory import Supermemory

client = Supermemory()

result = client.profile(container_tag="user_123")

print(result.profile.static)   # Long-term facts
print(result.profile.dynamic)  # Recent context
```

cURL:bash curl -X POST " \ -H "Authorization: Bearer $SUPERMEMORY_API_KEY" \ -H "Content-Type: application/json" \ -d '{"containerTag": "user_123"}'

Response:

{
  "profile": {
    "static": [
      "User is a software engineer",
      "User specializes in Python and React",
      "User prefers dark mode interfaces"
    ],
    "dynamic": [
      "User is working on Project Alpha",
      "User recently started learning Rust",
      "User is debugging authentication issues"
    ]
  }
}

---

Profile + Search

Get profile and search results in one call by adding the q parameter:

TypeScript: ```typescript const result = await client.profile({ containerTag: "user_123", q: "deployment errors" });

// Profile data
const { static: facts, dynamic: context } = result.profile;

// Search results (only if q was provided)
const memories = result.searchResults?.results || [];
```

Python: ```python result = client.profile( container_tag="user_123", q="deployment errors" )

# Profile data
facts = result.profile.static
context = result.profile.dynamic

# Search results
memories = result.search_results.results if result.search_results else []
```

---

Parameters

Parameter	Type	Required	Description
containerTag	string	Yes	User/project identifier
q	string	No	Search query (includes search results in response)
threshold	0-1	No	Filter search results by relevance score

---

Building Prompts

The most common pattern — inject profile into your LLM's system prompt:

async function chat(userId: string, message: string) {
  const { profile } = await client.profile({ containerTag: userId });

  const systemPrompt = `You are assisting a user.

ABOUT THE USER:
${profile.static?.join('\n') || 'No profile yet.'}

CURRENT CONTEXT:
${profile.dynamic?.join('\n') || 'No recent activity.'}

Personalize responses to their expertise and preferences.`;

  return llm.chat({
    messages: [
      { role: "system", content: systemPrompt },
      { role: "user", content: message }
    ]
  });
}

---

Full Context Pattern

Get profile + query-specific memories in one call:

async function getContext(userId: string, query: string) {
  const result = await client.profile({
    containerTag: userId,
    q: query,
    threshold: 0.6
  });

  return `
User Background:
${result.profile.static.join('\n')}

Current Context:
${result.profile.dynamic.join('\n')}

Relevant Memories:
${result.searchResults?.results.map(m => m.memory).join('\n') || 'None'}
  `;
}

---

Framework Examples

Express.js Middleware

async function withProfile(req, res, next) {
  if (!req.user?.id) return next();

  try {
    const { profile } = await client.profile({
      containerTag: req.user.id
    });
    req.userProfile = profile;
  } catch (e) {
    req.userProfile = null;
  }
  next();
}

app.use(withProfile);

app.post('/chat', (req, res) => {
  // req.userProfile available in all routes
});

Next.js API Route

// app/api/chat/route.ts
export async function POST(req: NextRequest) {
  const { userId, message } = await req.json();

  const { profile } = await client.profile({
    containerTag: userId
  });

  const response = await generateResponse(message, profile);
  return NextResponse.json({ response });
}

AI SDK Integration

import { withSupermemory } from "@supermemory/tools/ai-sdk"
import { openai } from "@ai-sdk/openai"

// Profiles automatically injected
const model = withSupermemory(openai("gpt-4"), {
  containerTag: "user-123",
  customId: "conv-1",
})

const result = await generateText({
  model,
  messages: [{ role: "user", content: "Help with my project" }]
});

See AI SDK Integration for details.

---

Response Schema

interface ProfileResponse {
  profile: {
    static: string[];   // Long-term facts
    dynamic: string[];  // Recent context
  };
  searchResults?: {     // Only if q parameter provided
    results: SearchResult[];
    total: number;
    timing: number;
  };
}

---

Next Steps

• User Profiles Concept — Understand static vs dynamic
• Ingesting Content — Build profiles by adding content
• AI SDK Integration — Automatic profile injection