Maniac: Your best model in one click.

Maniac is an enterprise AI platform that makes it easy to replace existing LLM API calls with fine-tuned, task-specific models. Drop Maniac in with one line of code to:

• Capture and structure production LLM traffic

• Automatically fine-tune and evaluate Small Language Models (SLMs) on your tasks

• Replace over-generalized LLM calls with higher performance, lower latency models built for just what you need

• Focus engineering time where it matters most: building and refining high-quality model evaluations—not managing infrastructure, hyperparameters, or bespoke fine-tuning pipelines

All with virtually no changes to your existing codebase.

Getting started

1

Sign up for Maniac

Head over to https://app.maniac.ai/auth/register

2

Create a new Organization

Organizations house multiple projects.

3

Add a Project

All your work — containers, evals, and deployments — live here.

4

Generate an API key

From your project settings

---

Dropping Maniac into your Codebase

Install the library

pip install maniac

Initialize client

from maniac import Maniac

# Simple initialization - Maniac handles all providers automatically
maniac = Maniac(api_key="your-maniac-api-key")

Create a container

Containers log inference and automatically build datasets for fine-tuning and evaluation. initial_model sets the model used in that container until a Maniac model is deployed.

container = maniac.containers.create(
  label = "my-container"
  initial_model = "openai/gpt-5",
  initial_system_prompt = "You are a helpful math tutor."
)

Run inference in a container

Running inference will auto-generate inference logs. Data can also be manually uploaded.

Chat Completions API

response = maniac.chat.completions.create(
    model = "maniac:my-container",
    messages = [{"role": "user", "content": "A train travels 120 miles in 2 hours. What is its average speed?"}],
    judge_prompt = "Compare two math solutions. Is A better than B? Consider: calculation accuracy, clear explanations, educational value."
    reasoning = {"effort": "medium"} # Optional reasoning parameter
)

print(response["choices"][0]["message"]["content"])
# Output: "The average speed is 60 miles per hour. This is calculated by dividing distance (120 miles) by time (2 hours): 120 ÷ 2 = 60 mph."

Note: We recommend defining the system prompt at the container level. All inference requests executed through that container will automatically inherit this system prompt. If a request’s messages array includes its own system prompt, it will override the container-level system prompt for that request only.

---

Optimizing your model

The inference logs in your container now serve as training data for a new SLM—fully yours, lower latency, cheaper, and optimized specifically for your task.

1

Create an Eval

Evaluations define the optimization target. They can be implemented as arbitrary code or defined using judge prompts.

From the Evals tab inside a container, Add Eval.

Judge Eval

Code Eval

2

Launch Optimization

Once you've defined an eval, the Optimization dashboard lets you configure and run post-training pipelines using techniques such as SFT, GRPO, and GEPA.

Each stage of the pipeline is modular, allowing you to select base models, the evaluation to optimize against, adjust hyperparameters, swap classifier heads, and experiment with different training strategies.

---

Deploy.

Optimized models can be be deployed into a container from the Models tab. Once deployed, you can chat with your generated models, and inference requests are now routed through the Maniac model instead of the initial_model.

Need help?

📧 Email us at [email protected]

We'll get back to you within a day.