Help Center

Faultline is an AI-powered claim verification engine that detects misinformation, validates sources, and assigns confidence scores to claims in real time. It supports text, URL, and document-based analysis.

Faultline can verify factual claims, statistical assertions, historical dates, scientific statements, and quoted attributions. It cross-references multiple source databases and assigns a confidence score from 0 to 100.

Each claim receives a confidence score (0–100) based on source reliability, corroboration count, recency of evidence, and logical consistency. Scores above 80 are considered high confidence, 50–80 moderate, and below 50 low confidence.

Yes. Faultline exposes a REST API for claim submission and result retrieval. You can submit claims programmatically and receive structured JSON responses with scores, sources, and reasoning traces.

Forge is a developer empowerment platform that provides AI-powered code generation, architecture planning, automated testing, and governance tools. It integrates into your CLI and IDE to accelerate development workflows.

Forge includes specialized agents for planning, building, testing, security scanning, performance optimization, database work, UI development, documentation, DevOps, and more. Each agent has domain-specific expertise and tools.

Forge uses a governance system with directives, CRUCIBLE test quality audits, and scope validation. The Oracle agent monitors changes in real time, and the CEO Loop handles autonomous strategic decisions.

Yes. The Forge testing agent generates unit, integration, and end-to-end tests. It analyzes your codebase patterns and creates tests that follow your existing conventions, including proper mocking and assertion styles.

SynApps is a composable AI application framework that lets you build, deploy, and orchestrate multi-agent workflows. It provides pre-built components for common AI patterns like RAG, classification, and data pipelines.

SynApps is model-agnostic and supports Claude, GPT-4, Gemini, Llama, and custom models. You can mix models within a single workflow, using different models for different steps based on cost and capability requirements.

Yes. SynApps includes a deployment pipeline with staging and production environments, health monitoring, automatic scaling, and rollback capabilities. Workflows can be deployed to Vercel, AWS, or self-hosted infrastructure.

Visit the Faultline product page and click "Try Demo" to test claim verification instantly. For API access, sign up for an account and generate your API key from the dashboard. Documentation is available at /docs/faultline.

Navigate to the Faultline demo page, paste or type a claim into the input field, and click "Verify." The system will analyze the claim and return a confidence score with supporting sources within seconds.

Run "/enable-forge" in your Claude Code CLI session to activate the Forge command center. Forge initializes automatically by reading your project structure and presenting a 4-option menu: Resume, Plan, Soundboard, or Health Check.

Use the Planner agent to design your feature architecture, then approve the plan and hand off to the Builder agent for implementation. The Guardian agent runs quality checks automatically after implementation completes.

Start by installing the SynApps SDK with "npm install @nxtg/synapps". Create a workflow file defining your steps, connect your AI model credentials, and run "synapps dev" to test locally before deploying.

SynApps provides connectors for PostgreSQL, MongoDB, Redis, S3, and REST APIs. Use the DataSource component to define your connection and the Pipeline component to process data through your AI workflow.

Low scores typically indicate limited source corroboration, ambiguous phrasing, or claims about very recent events where indexed sources are sparse. Try rephrasing the claim more specifically or providing additional context.

A 429 status indicates you have exceeded your rate limit. Free-tier accounts are limited to 100 requests per hour. Upgrade your plan or implement exponential backoff in your client.

Ensure you have activated Forge with "/enable-forge" in your current session. Check that your CLAUDE.md file exists and contains the required governance sections. If agents still fail, try restarting your CLI session.

CRUCIBLE Gate audits use static analysis heuristics that may flag intentional patterns as issues. Review each finding and use inline annotations to suppress known false positives. Update your test maturity profile if needed.

Workflow timeouts usually indicate a step is waiting on a slow external service or processing too much data. Check individual step durations in the monitoring dashboard, increase timeout limits, or add pagination to data-heavy steps.

Verify your connection string format and ensure your database allows connections from your deployment environment. Check firewall rules, SSL certificate configuration, and that the database user has the required permissions.