Getting Started

Sign in to MechaMental

Navigate to your MechaMental instance URL and sign in with your credentials. After authentication, you'll land on the workspace dashboard.

Create a Workspace

If your organization doesn't have a workspace yet, create one:

1. Go to Settings in the sidebar
2. Under Workspaces, click Create Workspace
3. Enter a name and optional description
4. Click Create

Create an App

Apps are the primary unit of work in MechaMental. Each app contains pipelines, endpoints, releases, and configuration.

1. Click Apps in the sidebar
2. Click the New App button
3. In the Create New App dialog, fill in:
   • App Name (required) — e.g., "Customer Support Agent"
   • Description (optional) — a short summary of what this app does
4. Click Create

You'll be taken to the app detail view, which has the following tabs: Dashboard, Namespaces, Augmentations, Environments, Tokens, Releases, and Settings.

Create an Augmentation Pipeline

Augmentations are the AI pipelines that process requests.

1. Open the Augmentations tab in your app
2. Click New Pipeline
3. Enter a name (e.g., "Support Chat Pipeline")
4. The pipeline editor opens with an empty canvas

Add a Stage and Inference Step

The pipeline editor displays stages horizontally from left to right, each rendered as a card. Steps within a stage are listed vertically.

1. Click Add Stage on the canvas
2. In the Create Stage dialog, enter a Name (e.g., "Main") and an optional Description
3. Inside the new stage card, click Add Step
4. From the step type dropdown, select inference
5. Give the step a name (e.g., "chat_response")
6. Select a model from the available logical models

Configure the Prompt

1. Click on your inference step to open the step detail dialog
2. Navigate to the Prompt tab (available only for inference steps)
3. Write a system prompt using Jinja2 template syntax:

You are a helpful customer support assistant.

The user's question: {{ endpoint.payload.query }}

Respond concisely and professionally.

Create an Endpoint

Endpoints expose your pipeline as an HTTP API.

1. In the pipeline editor top bar, click Create Endpoint
2. Fill in the endpoint fields:
   • Name — e.g., "Chat Endpoint"
   • Path — e.g., /chat
   • Description — optional
   • HTTP Method — typically POST
   • Endpoint Type — select the appropriate type
3. Optionally define an input schema to validate incoming payloads
4. Save the endpoint
5. After creation, generate an API key for authentication

Test via curl

Send a test request to your new endpoint:

curl -X POST https://api.mechamental.com/v1/trigger/chat \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "How do I reset my password?"}'

You should receive a response generated by the LLM using your configured prompt.

Create a Release and Deploy

When your pipeline is working correctly, package it for deployment:

1. Go to the Releases tab in your app
2. Click Create Release
3. In the Create Release dialog, enter:
   • Release Tag — e.g., v1.0.0
   • Description — e.g., "Initial release of support chat"
4. Click Create — the release is created in the Ready state
5. Click Deploy to Environment and select an environment (e.g., Production)
6. The release moves to the Active state

Your pipeline is now live and serving traffic in the selected environment.