SOVEREIGN
INTEGRATED TESTING ENVIRONMENT
Replacing fragmented testing tools with A UNIFIED LOCAL WORKSPACE.
Write. Run. Heal. Secure.
SOVEREIGN is a Beta release. Its AI features run on your key (Bring Your Own Key): you supply an API key for DeepSeek, Gemini, or OpenAI, and SOVEREIGN calls that provider directly from your machine. Being straight about the trade-off: the prompt content for those features does go to the provider you chose — we never see it, we never proxy it, and there is no hyenai server in the path. Everything else — test cases, credentials, execution results, reports — stays on disk.
Need zero egress? Point the AI Hub at Ollama and SOVEREIGN makes no outbound network request at all. That is the configuration to use for air-gapped work. Full data-flow disclosure →
DASHBOARD
The Command Center. High-performance metrics aggregator pulling real-time statistics from the local Execution Engine. Features trend heatmaps and immediate status reports.
[TC] TEST EDITOR
The Architect. A dense local IDE. Governs the TestCaseItem data model with offline CRUD operations and AI-assisted expansion of steps via the local reasoning service.
[AI] AUTOMATION AGENT
The Automator. The heavy-lifting execution core. Leverages a local BrowserContextPool to lease Playwright instances. Translates natural language into JSON actions.
[AP] API TESTING
The Communicator. High-speed REST interaction layer. Features an auto-token extraction system that detects authentication handshakes to automate subsequent requests.
[VS] STORYSIGHT
The Analyst. Scores user stories for defect risk using a locally-trained ML.NET FastTree classifier, then lints them for the ambiguities that make a story untestable (an upload with no file named, a search with no term, a navigation with no URL) and drafts a page object from the validated DOM. The model runs on your machine; no story text is sent anywhere unless you enable a BYOK provider.
AI HUB
The Oracle. Central routing point for local AI interactions. Manages multiple providers (DeepSeek, ChatGPT, Gemini, Ollama) with seamless auto-failover and aggressive caching.
[LT] LOAD TESTING
The Tank. Stress-testing module designed to simulate high-concurrency scenarios locally. Spawns threads to measure system threshold.
[RT] RED TEAM AGENT
The Adversary. Active-exploit engine. Features a Red Team Agent for fuzzed verification (SQLi, XSS) and a local audit engine for credentials scanning.
[HH] HEALING HISTORY
The Repair Ledger. Audit trail for locator repairs. Every proposed change to an element definition is recorded with its old selector, the replacement, and who approved it — so a repair can be reviewed or reverted. Automatic repair during execution is in development and not active in v0.9; today this screen shows the approval history for repairs made through the object repository.
ROI ANALYTICS
The Calculator. ROI analysis. Translates execution runs, automated hours, and caught defects into financial impact metrics and velocity charts.
[ST] STRATEGY PLANNER
The Planner. Connects high-level testing plans with specific execution runs. Manages timelines and builds to provide insight into release readiness.
VERSION CONTROL
The Versioner. Direct Git integration. Treats Test Cases as Code, allowing for commits, branches, and merges directly within the workspace.
ADAPTIVE_TIMEOUTS
Intelligent processing windows designed for complex reasoning and deep-dive analysis without premature timeout.
IDEMPOTENT_TASKS
Self-healing tasks capable of indefinite execution loops without state corruption or data loss.
SILENT_FAILURES
Aggressive error detection providing immediate, actionable feedback upon any system discrepancy.
SOVEREIGN AGENTIC SYSTEM
SOVEREIGN runs as a desktop application inside your network. Test logic, credentials, and results are stored locally and encrypted at rest with Windows DPAPI — there is no hyenai account, no sync service, and no telemetry. The one exception is deliberate and disclosed: AI features call the provider whose key you supply. Choose Ollama and that exception disappears entirely.
INDUSTRIAL ENGINE PRECISION
Engineered for endurance. Our FAULT-TOLERANT technology isolates execution threads. WORKSPACE GUARD protects against erratic input and adversarial script injection.
DEPLOY LOCAL ENVIRONMENT
VERSION V0.9 // WIN64 // SOVEREIGN EDITION