API Documentation

Everything you need to integrate Seizn into your applications. Add persistent memory to your AI with just a few lines of code.

Quick Start

Get your API key from the dashboard, then start making requests:

bash

# Add a memory
curl -X POST https://seizn.com/api/memories \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"content": "User prefers dark mode interfaces"}'

# Search memories
curl "https://seizn.com/api/memories?query=user+preferences" \
  -H "Authorization: Bearer YOUR_API_KEY"

Authentication

All API requests require an API key passed in the Authorization: Bearer header.

bash

curl -H "Authorization: Bearer szn_your_api_key_here" \
  https://seizn.com/api/memories?query=test

Security: Keep your API keys secret. Never expose them in client-side code. Use environment variables or a backend proxy.

API Endpoints

POST/api/memories

Add a new memory to the user's memory store.

Request Body

content-string (required)

memory_type-string - fact, preference, experience, relationship, instruction

tags-string[]

namespace-string (default: "default")

scope-string - user, session, agent

session_id-string

agent_id-string

Response

{
  "success": true,
  "memory": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "content": "User prefers dark mode interfaces",
    "memory_type": "preference",
    "tags": ["ui", "settings"],
    "namespace": "default",
    "created_at": "2026-01-08T10:30:00Z"
  }
}

GET/api/memories

Search memories using semantic similarity.

Query Parameters

query-string (required)

limit-number (default: 10, max: 100)

threshold-number 0-1 (default: 0.7)

namespace-string

Response

{
  "success": true,
  "results": [
    {
      "id": "550e8400-e29b-41d4-a716-446655440000",
      "content": "User prefers dark mode interfaces",
      "memory_type": "preference",
      "tags": ["ui", "settings"],
      "similarity": 0.89
    }
  ],
  "count": 1
}

DELETE/api/memories

Delete memories by their IDs.

Query Parameters

ids-string (required) - comma-separated

Response

{
  "success": true,
  "deleted": 3
}

POST/api/extract

Extract and store memories from a conversation using AI.

Request Body

conversation-string (required)

model-string - haiku | sonnet (default: haiku)

auto_store-boolean (default: true)

namespace-string (default: "default")

Response

{
  "message": "Extracted 3 memories, stored 3",
  "extracted": [
    {
      "content": "User is a software developer working with Python",
      "memory_type": "fact",
      "tags": ["profession", "programming"],
      "confidence": 0.95,
      "importance": 7
    }
  ],
  "stored": [...]
}

POST/api/query

Get AI-generated responses using relevant memories as context (RAG).

Request Body

query-string (required)

model-string - haiku | sonnet (default: haiku)

top_k-number (default: 5)

namespace-string

include_memories-boolean (default: true)

Response

{
  "response": "Based on your preferences, I'd recommend using VS Code with a dark theme...",
  "memories_used": [
    {
      "id": "550e8400-e29b-41d4-a716-446655440000",
      "content": "User prefers dark mode interfaces",
      "similarity": 0.85
    }
  ],
  "model_used": "haiku"
}

Rate Limits

Plan	Monthly API Calls	Max Memories	API Keys
Free	1,000	100	2
Starter	50,000	5,000	3
Plus	500,000	50,000	5
Pro	2,000,000	Unlimited	10
Enterprise	Unlimited	Unlimited	100

When you exceed your rate limit, the API returns a 429 Too Many Requests response.

Error Codes

Code	Description
`200`	Success
`400`	Bad Request - Missing or invalid parameters
`401`	Unauthorized - Invalid or missing API key
`429`	Too Many Requests - Rate limit exceeded
`500`	Internal Server Error - Something went wrong

Security & Governance

Data Security

✓Encryption at rest: All data encrypted with AES-256
✓Encryption in transit: TLS 1.3 for all connections
✓Tenant isolation: Complete data separation between accounts

API Key Management

✓Key rotation: Rotate keys anytime from the dashboard
✓Key expiration: Auto-expire keys after 90 days (configurable)
✓Usage tracking: Monitor per-key usage in real-time

Data Retention & Deletion

✓Export: Export all your data anytime via API or dashboard
✓Deletion: Hard delete with no retention after 30 days
✓GDPR/CCPA: Full compliance with data subject rights

SDKs

Python

bash

pip install seizn

python

from seizn import Seizn

client = Seizn(api_key="your_api_key")

# Add memory
client.add("User prefers dark mode")

# Search
results = client.search("preferences")

# Extract from conversation
client.extract(conversation="...")

JavaScript

bash

npm install seizn

javascript

import { Seizn } from 'seizn';

const client = new Seizn({ apiKey: 'your_api_key' });

// Add memory
await client.add('User prefers dark mode');

// Search
const results = await client.search('preferences');

// Extract from conversation
await client.extract({ conversation: '...' });

MCP Server — Every Editor, One Memory

The Seizn MCP Server (seizn-mcp) bridges your Seizn memories to AI coding assistants via the Model Context Protocol. 40+ tools, MCP Resources, webhooks, OAuth device flow, and multi-editor config sync — all in one package.

bash

# Install globally or use npx
npx seizn-mcp@latest

# Or add to Claude Code settings (~/.claude/settings.json)
{
  "mcpServers": {
    "seizn": {
      "command": "npx",
      "args": ["-y", "seizn-mcp@latest"],
      "env": { "SEIZN_API_KEY": "your-api-key" }
    }
  }
}

Supported Editors

> Claude Code — native MCP
> Cursor — native MCP
> Windsurf — native MCP
> Cline — native MCP
~ GitHub Copilot — via config sync
~ Aider — via config sync
~ OpenAI Codex — via config sync

Key Features

> 40+ MCP Tools — memories, knowledge graph, profile, webhooks, config sync
> MCP Resources — seizn://memories/recent, seizn://profile, seizn://context/{format}
> OAuth Device Flow — browser auth, no API key copy
> Auto Context — detects project from package.json, pyproject.toml, Cargo.toml
> UTF-8 Support — Korean, Japanese, Chinese, Arabic and 100+ languages

Multi-Editor Config Sync

Seizn exports your memories as editor-specific configuration files. Your AI preferences follow you across every tool.

File	AI Tool	Method
CLAUDE.md	Claude Code	MCP + File
AGENTS.md	OpenAI Codex	File Sync
.cursor/rules	Cursor	MCP + File
.windsurfrules	Windsurf	MCP + File
.github/copilot-instructions.md	GitHub Copilot	File Sync
.clinerules	Cline	MCP + File
CONVENTIONS.md	Aider	File Sync

OAuth Device Flow

No more copying API keys. The MCP server supports RFC 8628 Device Authorization Grant for browser-based authentication.

Run auth_login tool

Enter code ABCD-1234 in browser

Token saved to ~/.seizn/

Zero-copy auth: The device flow generates a human-readable code, opens your browser, and saves credentials automatically. Works with any terminal or SSH session.