Mandatum

Appearance

MandatumPrompt Engineering Platform

Manage, version, and execute AI prompts with custom code integration

Get Started
Mandatum
🔒

Your Keys Stay Local

Your LLM API keys never touch our servers. The SDK executes calls locally and only logs results back to Mandatum.

📝

Prompt Management

Create, edit, and delete prompts with a visual editor. Version control with git-like history tracks every change.

💻

Custom Code Integration

Write pre-run and post-run scripts in Python. Transform inputs before LLM calls and parse outputs after.

📊

Request Logging & Analytics

Every LLM request is automatically logged. Track costs, latency, and usage patterns with detailed analytics.

🔌

Multi-Provider Support

Use OpenAI, Anthropic, Google AI, and more. Switch providers without changing code.

🚀

Simple Integration

One SDK, one API key. Run prompts from your application with a single function call.

What is Mandatum? ​

Mandatum is a prompt engineering platform for teams who want full control over their LLM integrations. Your API keys stay on your system - they never touch our servers. Mandatum handles prompt management, versioning, logging, and analytics.

How It Works ​

Your App (with your LLM keys)
        ↓
    Mandatum SDK
        ↓
1. Fetches prompt template from Mandatum
2. Executes LLM call locally (your keys never leave your system)
3. Logs results back to Mandatum for analytics
  1. Sign up and get your Mandatum API key
  2. Create prompts in the dashboard with variables like
  3. Add custom code (optional) - pre-run scripts to transform input, post-run scripts to parse output
  4. Install the SDK and use your own LLM provider keys locally
  5. View logs and analytics - every request is tracked with cost, latency, and metadata

Key Features ​

FeatureDescription
PromptsCreate, edit, delete, and version prompt templates
IntegrationsWrite pre-run and post-run Python scripts that execute with your prompts
EvaluationsTest prompt quality, compare versions, run automated checks
LoggingAutomatic request logging with full input/output capture
AnalyticsCost tracking, latency monitoring, usage dashboards

Quick Example ​

python
from openai import OpenAI
from mandatum import Mandatum

# Your LLM keys stay local - never sent to Mandatum
openai_client = OpenAI(api_key="your-openai-key")
mandatum = Mandatum(api_key="your-mandatum-api-key")

response = mandatum.prompts.run(
    prompt_id="your-prompt-id",
    user_input="I need help with my order",
    llm_client=openai_client,
    variables={"customer_name": "Alice"},
    environment="production",
)

print(response.output)
print(f"Cost: ${response.cost:.4f}")

Ready to Get Started? ​