Learn How to Use Letaria

Sign up for Letaria using your email or OAuth providers (Google, GitHub). After registration, you'll be prompted to set up your first workspace.

**Steps:** 1. Click "Get Started" on the homepage 2. Choose your sign-up method 3. Verify your email address 4. Complete your profile setup

Workspaces help you organize projects and collaborate with team members. Your first workspace is created automatically.

**Workspace Settings:** • Name your workspace (e.g., "My Company QA") • Invite team members (optional) • Configure default settings

Projects contain your requirements, generated scenarios, and test cases. Create a project to get started.

**Steps:** 1. Click "New Project" in your workspace 2. Enter a project name and description 3. Choose your workflow (Standard or Advanced) 4. Start uploading requirements

Connect Letaria to your JIRA Cloud instance using OAuth 2.0 authentication.

**Requirements:** • JIRA Cloud account (Data Center not supported yet) • Admin or user permissions to create OAuth connections • JIRA project access

Once connected, you can import issues from any accessible project.

**Supported Issue Types:** • User Stories • Requirements • Tasks with acceptance criteria • Epics and sub-tasks

**Import Options:** • Select specific issues by ID • Use JQL queries for bulk import • Filter by labels, sprints, or versions

Maintain bidirectional links between JIRA issues and generated test cases.

**Features:** • Create remote links in JIRA • Track coverage per requirement • Update links when test cases change

Connect to your Azure DevOps organization using OAuth authentication.

**Requirements:** • Azure DevOps Services account • Access to the project you want to import from • Personal Access Token or OAuth consent

Import work items from Azure DevOps boards and backlogs.

**Supported Work Item Types:** • User Stories • Product Backlog Items • Requirements • Features

**Import Methods:** • Select items from a query • Browse by area path • Import by ID

Upload your requirements documents to start generation.

**Supported Formats:** • PDF documents • Microsoft Word (.docx) • Plain text (.txt) • Markdown (.md)

**Best Practices:** • Use clear, concise language • Include acceptance criteria • Separate functional from non-functional requirements

Our AI engine analyzes your requirements and generates comprehensive test scenarios.

**What Happens:** 1. Document parsing and chunking 2. Requirement extraction and analysis 3. Multi-agent scenario generation (6+ agents cross-validating) 4. Scenario generation in Gherkin format 5. Coverage analysis and gap detection

**Generation Options:** • Standard (balanced coverage) • Comprehensive (maximum coverage) • Focused (specific areas only)

Every generated scenario includes explainability data.

**EAI Features:** • **Citations** - Links to source requirements • **Rationale** - Why this scenario was generated • **Confidence Score** - AI certainty (0-100%) • **Assumptions** - Inferences made by AI • **Questions** - Areas needing clarification

Review AI-generated scenarios before approval.

**Review Features:** • Side-by-side comparison with requirements • Inline editing and comments • Version comparison • Batch review mode

Configure approval workflows based on your governance needs.

**Approval Policies:** • **Any** - One approver is sufficient • **All** - All reviewers must approve • **Quorum** - Majority approval required

**Digital Signatures:** • Timestamp and user tracking • Audit trail for compliance • Export approval records

Provide feedback and regenerate scenarios as needed.

**Options:** • Regenerate with additional context • Apply feedback to future generations • Lock approved scenarios from changes

Export your test cases in multiple formats.

**Available Formats:** • **Excel (.xlsx)** - Full formatting with metadata • **CSV** - Simple comma-separated values • **JSON** - For API integration

**Export Options:** • Include/exclude EAI metadata • Filter by status or approval • Select specific scenarios

Import exported files into your test management tools.

**Compatible With:** • TestRail (via CSV/Excel) • Xray for JIRA (via JSON) • qTest (via Excel) • Zephyr (via Excel) • And many more...

Workspaces provide isolation and organization for your projects.

**Features:** • Separate projects by team or department • Isolated integrations per workspace • Workspace-specific settings

Invite team members to collaborate in your workspace.

**Roles:** • **Owner** - Full control, billing access • **Admin** - Manage members, projects • **Member** - Create and edit content • **Viewer** - Read-only access

Letaria uses cloud orchestration with on-premises execution agents. The cloud layer (Letaria Cloud) orchestrates jobs, while execution agents run tests inside your infrastructure.

Key Concepts: • Cloud Orchestration — Job routing, scheduling, result aggregation (Letaria Cloud) • Execution Agents — AI-powered agents that perform test actions (your infrastructure) • Agent Types — Specialized agents for different platforms

Letaria supports specialized execution agents for a wide range of enterprise platforms.

Available Now: • Web Browser Execution Agent — Playwright/Chromium for web application testing

Coming Soon: • API Execution Agent — REST and GraphQL endpoint testing

On Roadmap: • Windows Desktop (.NET/WPF/WinForms) • SAP GUI Scripting • Mainframe Terminals (3270/5250/AS400) • Citrix/VDI (image-based) • Mobile (iOS/Android via Appium) • Java Desktop (Swing/AWT) • Oracle Forms

Execution agent deployment options depend on your plan tier.

Cloud Execution Agents (Pro & Team plans): • Managed cloud infrastructure (AWS, Azure, GCP) • Serverless or container-based deployment • No infrastructure management required • Automatic scaling

On-Premises Execution Agents (Enterprise only): • Deploy inside your corporate network, VPC, or air-gapped environments • All test data stays within your infrastructure • Encrypted connectivity to Letaria Cloud Orchestration • Full control over agent configuration and access

Every test execution generates professional reports with evidence.

Report Features: • HTML reports with screenshot evidence • Pass/fail summaries with execution duration • Compressed screenshot attachments (PNG to JPG) • Collapsible test case details • Exportable audit data

Unlike single-model AI testing tools, Letaria deploys 6+ specialized AI agents that cross-validate every decision before execution.

Why Multi-Agent: • Single AI models can hallucinate selectors or actions • Cross-validation catches errors before execution • 61.2% more efficient than single-model approaches • 33% faster execution through specialized agents

Each agent has a specialized role in the validation pipeline.

Agent Pipeline: 1. Step Interpreter — Parses natural language test steps into structured actions with intent classification 2. Element Locator — AI-driven element identification with confidence scoring and alternative candidates 3. Input Guard — Sanitizes all inputs before they reach the LLM; blocks prompt injection and strips scripts 4. Output Validator — Validates every LLM response for dangerous patterns, sensitive data leakage, and confidence thresholds 5. Consistency Checker — Cross-validates AI reasoning against actions; detects contradictions and applies confidence penalties 6. Circuit Breaker — Halts execution on consecutive failures; prevents cascading errors and wasted tokens

Bring Your Own Keys mode lets you use your own LLM API keys for AI processing.

How It Works: • Configure your own API keys for OpenAI, Anthropic Claude, Google Gemini, or other providers • Your data is processed through your accounts • Never used for third-party model training • Available as overflow when exceeding plan limits (pay LLM provider cost only, no Letaria markup)

Enterprise BYOK: • Full BYOK as primary mode (not just overflow) • Data sovereignty and cost management control • Compatible with on-premises execution agent deployments