Skip to main content
Memori Advanced Augmentation uses a quota system to manage memory storage and API usage. Understanding and managing your quota ensures uninterrupted service and optimal performance.

​
Understanding Quotas

​
What Counts Toward Your Quota

Your quota tracks the number of memories stored across all entities, processes, and sessions:
  • Conversation turns - Each LLM interaction
  • Augmented memories - Extracted facts, preferences, events, etc.
  • Session data - Grouped interactions
  • Embeddings - Vector representations for semantic search

​
Quota Tiers

Memori provides different quota tiers based on your authentication method:
TierAuthenticationMax MemoriesBest For
IP-BasedNone (anonymous)1,000Testing, evaluation
Free DeveloperAPI key100,000Development, small projects
EnterpriseCustomUnlimitedProduction deployments
Memori Advanced Augmentation is always free for developers. Sign up for an API key to increase your quota from 1,000 to 100,000 memories.

​
Checking Your Quota

​
Using the CLI

The fastest way to check your quota: 
python -m memori quota
Example Output: 
 __  __                           _ 
|  \/  | ___ _ __ ___   ___  _ __(_)
| |\/| |/ _ \ '_ ` _ \ / _ \| '__| |
| |  | |  __/ | | | | | (_) | |  | |
|_|  |_|\___|_| |_| |_|\___/|_|  |_|
                  perfectam memoriam
                       memorilabs.ai
                            v3.2.1

+ Maximum # of Memories: 100,000
+ Current # of Memories: 45,678

+ You are currently using 45.68% of your quota.

​
Using the Dashboard

Visit app.memorilabs.ai to:
  • View real-time quota usage
  • Browse all stored memories
  • Analyze memory distribution by entity and process
  • Track usage trends over time
  • Export memory data

​
Programmatic Quota Checking

You can check quota programmatically by calling the Memori API: 
import os
import requests

api_key = os.getenv("MEMORI_API_KEY")
headers = {"Authorization": f"Bearer {api_key}"}

response = requests.get(
    "https://api.memorilabs.ai/v1/sdk/quota",
    headers=headers
)

quota = response.json()
print(f"Using {quota['memories']['num']} of {quota['memories']['max']} memories")
print(f"Percentage: {(quota['memories']['num'] / quota['memories']['max']) * 100:.2f}%")

​
Quota Optimization Strategies

​
1. Efficient Entity and Process Attribution

Proper attribution prevents duplicate memories and improves organization: 
from memori import Memori
from openai import OpenAI

mem = Memori()
client = mem.llm.register(OpenAI())

# Good: Specific attribution per user and process
mem.attribution(
    entity_id=f"user_{user.id}",
    process_id="customer_support_agent"
)

# Avoid: Generic attribution creates noise
mem.attribution(
    entity_id="default",
    process_id="main"
)
Best Practice: Use unique entity IDs for each user and descriptive process IDs for different agent types. This improves memory recall accuracy and makes quota usage more transparent.

​
2. Session Management

Group related interactions into sessions to optimize memory storage: 
# Start a new session for a distinct conversation
mem.new_session()

# Process multiple related interactions
for message in conversation:
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": message}]
    )

# Explicitly start a new session for a new topic
mem.new_session()

​
3. Recall Limits and Thresholds

Configure recall settings to balance performance and quota usage: 
from memori import Memori

mem = Memori()

# Adjust recall limits
mem.config.recall_embeddings_limit = 500  # Default: 1000
mem.config.recall_facts_limit = 3          # Default: 5
mem.config.recall_relevance_threshold = 0.2  # Default: 0.1
  • recall_embeddings_limit: Maximum number of embeddings to search during recall. Lower values = faster queries, potentially less accurate recall.
  • recall_facts_limit: Maximum number of facts to include in augmentation. Lower values = less context, faster processing.
  • recall_relevance_threshold: Minimum similarity score (0-1) for memories to be recalled. Higher values = more selective recall.

​
4. Environment Variables for Global Configuration

Set quota-related configuration via environment variables: 
# Reduce embeddings search space (lower quota usage)
export MEMORI_RECALL_EMBEDDINGS_LIMIT=500

# Use a more efficient embedding model
export MEMORI_EMBEDDINGS_MODEL="all-MiniLM-L6-v2"

​
Monitoring Quota Usage

​
Set Up Alerts

Create a monitoring script to alert when approaching quota limits: 
import os
import requests
from datetime import datetime

def check_quota_and_alert(threshold=0.9):
    """Alert if quota usage exceeds threshold (default 90%)"""
    api_key = os.getenv("MEMORI_API_KEY")
    headers = {"Authorization": f"Bearer {api_key}"}
    
    response = requests.get(
        "https://api.memorilabs.ai/v1/sdk/quota",
        headers=headers
    )
    
    quota = response.json()
    usage = quota['memories']['num'] / quota['memories']['max']
    
    if usage >= threshold:
        print(f"⚠️  WARNING: Quota at {usage*100:.1f}%")
        print(f"   Using {quota['memories']['num']:,} of {quota['memories']['max']:,} memories")
        print(f"   Timestamp: {datetime.now().isoformat()}")
        # Send alert (email, Slack, etc.)
        return True
    
    print(f"✓ Quota OK: {usage*100:.1f}%")
    return False

if __name__ == "__main__":
    check_quota_and_alert(threshold=0.9)

​
Integrate with CI/CD

Add quota checks to your deployment pipeline: 
# .github/workflows/deploy.yml
name: Deploy

on: [push]

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      
      - name: Check Memori Quota
        env:
          MEMORI_API_KEY: ${{ secrets.MEMORI_API_KEY }}
        run: |
          pip install memori requests
          python -m memori quota
          python scripts/check_quota.py  # Custom quota check script

​
Handling Quota Limits

​
What Happens When You Reach Your Quota

When you reach your quota limit:
  1. Existing memories continue to be recalled normally
  2. New memories are not created (no errors, graceful degradation)
  3. You receive email notification (for API key holders)
  4. LLM interactions continue but without memory augmentation
Memori will not throw errors when quota is exceeded. Instead, it gracefully degrades to operate without creating new memories. Monitor your quota to avoid unexpected behavior.

​
Increasing Your Quota

1

Sign up for an API key

If you’re using IP-based quota, sign up for a free developer account:
python -m memori sign-up [email protected]
This increases your quota from 1,000 to 100,000 memories.
2

Request enterprise quota

For production deployments requiring >100,000 memories, contact the Memori team:
3

Use BYODB for unlimited storage

Deploy Memori BYODB (Bring Your Own Database) for unlimited memory storage:
from memori import Memori
import psycopg2

# Use your own PostgreSQL database
connection = psycopg2.connect(
    host="your-db-host",
    database="memori",
    user="your-user",
    password="your-password"
)

mem = Memori().storage.register(connection)
# No quota limits with BYODB!
See BYODB Documentation for details.

​
Quota Best Practices

​
Do’s

✅ Use specific entity and process IDs - Better organization, easier monitoring ✅ Implement session management - Group related interactions efficiently ✅ Monitor quota regularly - Set up automated alerts at 80-90% usage ✅ Test with IP-based quota - Validate before committing to production ✅ Consider BYODB for production - Unlimited storage, full control

​
Don’ts

❌ Don’t use generic IDs - Avoid entity_id="user" or process_id="agent" ❌ Don’t ignore quota warnings - Plan ahead before hitting limits ❌ Don’t create unnecessary sessions - Each session consumes quota ❌ Don’t forget to claim CockroachDB clusters - 7-day expiration for unclaimed clusters

​
Troubleshooting Quota Issues

​
Quota Not Updating

Problem: The quota command shows outdated numbers.Solutions:
  1. Quota updates may have a brief delay (typically less than 1 minute)
  2. Check the dashboard for real-time data: app.memorilabs.ai
  3. Verify you’re checking the correct account: 
    echo $MEMORI_API_KEY
python -m memori quota

​
Unexpected High Usage

Problem: Using more memories than anticipated.Investigation:
  1. Check the dashboard to see memory distribution by entity/process
  2. Look for generic IDs that might be capturing too much
  3. Review session management - are you creating too many sessions?
  4. Verify attribution is set correctly: 
    # Log your attributions
print(f"Entity: {mem.config.entity_id}")
print(f"Process: {mem.config.process_id}")
print(f"Session: {mem.config.session_id}")

​
API Key Not Increasing Quota

Problem: API key doesn’t seem to increase quota.Solutions:
  1. Verify API key is set correctly: 
    echo $MEMORI_API_KEY
  2. Ensure API key has no extra whitespace: 
    export MEMORI_API_KEY=$(echo $MEMORI_API_KEY | tr -d '[:space:]')
  3. Check for typos in the environment variable name
  4. Restart your application after setting the key
  5. Verify API key is valid: 
    python -m memori quota

​
Next Steps

CLI Usage

Learn all CLI commands for quota management

Performance Tuning

Optimize memory recall and quota efficiency

Dashboard

Monitor your quota in real-time

BYODB Setup

Deploy with unlimited storage

Build docs developers (and LLMs) love

Get started for freeTalk to us