Skip to main content

​
Overview

Without sessions, you manage conversation history manually by sending all prior messages on every request. Sessions do this for you server-side, stored in Redis with a 7-day TTL.
Without sessionsWith sessions
You send 50 messages of history every requestYou send only the new message
You manage message arrays in your appNinjaChat stores and injects history
Large request payloadsMinimal payload
History lost on app restart7-day persistent storage

​
Create a session

POST https://ninjachat.ai/api/v1/sessions
Authorization: Bearer nj_sk_YOUR_API_KEY
Content-Type: application/json

{ "session_id": "user-alice-session-1" }
Response: 
{
  "session_id": "user-alice-session-1",
  "message_count": 0,
  "created_at": "2026-03-10T12:00:00.000Z"
}
You can also omit session_id to get an auto-generated one: 
{}

{
  "session_id": "sess_a1b2c3d4e5f6g7h8i9j0k1l2",
  "message_count": 0,
  "created_at": "2026-03-10T12:00:00.000Z"
}

​
Session ID format

session_id must match: ^[a-zA-Z0-9_-]{1,64}$ — alphanumeric, hyphens, and underscores, 1–64 characters. Good session IDs: user-123, conv_abc, proj-2026-march, thread_def456

​
Chat with a session

Add session_id to any /chat request. Previous messages are automatically injected before your new messages. 
curl -X POST https://ninjachat.ai/api/v1/chat \
  -H "Authorization: Bearer nj_sk_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5",
    "messages": [{"role": "user", "content": "My name is Alice. I work in fintech."}],
    "session_id": "user-alice-session-1"
  }'
Second request (history is automatically included): 
curl -X POST https://ninjachat.ai/api/v1/chat \
  -H "Authorization: Bearer nj_sk_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5",
    "messages": [{"role": "user", "content": "What do you know about me?"}],
    "session_id": "user-alice-session-1"
  }'
The model will respond: “You told me your name is Alice and you work in fintech.”

​
Chat response with session info

{
  "model": "gpt-5",
  "choices": [{"message": {"role": "assistant", "content": "You told me your name is Alice..."}}],
  "session": {
    "id": "user-alice-session-1",
    "message_count": 4
  },
  "cost": {"this_request": "$0.006"}
}

​
Retrieve a session

GET https://ninjachat.ai/api/v1/sessions/user-alice-session-1
Authorization: Bearer nj_sk_YOUR_API_KEY

{
  "session_id": "user-alice-session-1",
  "messages": [
    {"role": "user", "content": "My name is Alice. I work in fintech."},
    {"role": "assistant", "content": "Nice to meet you, Alice! ..."},
    {"role": "user", "content": "What do you know about me?"},
    {"role": "assistant", "content": "You told me your name is Alice..."}
  ],
  "message_count": 4,
  "created_at": "2026-03-10T12:00:00.000Z",
  "updated_at": "2026-03-10T12:01:32.000Z"
}

​
Export a session

Export as JSON or Markdown — useful for saving conversations, support tickets, or debugging. 
# JSON export
GET https://ninjachat.ai/api/v1/sessions/user-alice-session-1/export?format=json
Authorization: Bearer nj_sk_YOUR_API_KEY

# Markdown export
GET https://ninjachat.ai/api/v1/sessions/user-alice-session-1/export?format=markdown
Authorization: Bearer nj_sk_YOUR_API_KEY
The Markdown export returns a formatted document with headers and message blocks, ready to save as a .md file.

​
Delete a session

DELETE https://ninjachat.ai/api/v1/sessions/user-alice-session-1
Authorization: Bearer nj_sk_YOUR_API_KEY

{ "deleted": true }

​
Limits and behavior

PropertyValue
TTL7 days from last activity
Max messages stored100
TruncationWhen limit is hit, oldest 50 messages are dropped
OwnershipSessions are tied to your API key — other keys cannot access them
StorageUpstash Redis (global, low latency)

​
Code examples

import requests, os
import uuid

API_KEY = os.environ["NINJACHAT_API_KEY"]
BASE = "https://ninjachat.ai/api/v1"
HEADERS = {"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"}

def create_session(user_id: str) -> str:
    """Create or reuse a session for this user."""
    session_id = f"support-{user_id}"
    requests.post(f"{BASE}/sessions", headers=HEADERS, json={"session_id": session_id})
    return session_id

def chat(session_id: str, message: str) -> str:
    """Send a message and get a reply — history is automatic."""
    r = requests.post(f"{BASE}/chat", headers=HEADERS, json={
        "model": "gpt-5",
        "messages": [{"role": "user", "content": message}],
        "session_id": session_id,
    })
    return r.json()["choices"][0]["message"]["content"]

# Usage
session = create_session("alice-123")
print(chat(session, "Hi, my order #4521 hasn't arrived yet."))
print(chat(session, "I ordered it 2 weeks ago."))
print(chat(session, "What's the refund policy?"))
# Each message includes full prior context automatically
Cache is automatically disabled when session_id is present — since history changes every request, cached responses would be incorrect.