Yes, you can upload a file (PDF, CSV, DOCX, etc), paste a URL, or write directly in the editor. Cloud storage integrations (AWS, Google Drive) are coming soon. In "Studio", click on "Guardrails", and under Input section, toggle on "Custom Rules", then select "Upload Rule" for uploading or pasting a URL, or "Text Rule" for writing directly.

Yes. Every team gets access to documentation, sample policies, and quick-start guides to help you get set up fast. Standard Plan customers also receive priority onboarding support, so your team can move from testing to production with confidence.

Yes. The Standard Plan includes generous usage, and anything beyond that scales on a pay-as-you-go basis. That means you can keep growing, adding more tokens, policy requests, projects, RAG files, or team members, without hitting hard limits or needing to switch plans.

Yes. For larger teams or regulated industries, we offer enterprise and custom pricing with higher usage limits, dedicated support, and tailored deployment options. Contact us to discuss a plan that fits your organization's needs.

You decide. CLoD offers three modes: store raw (full logs for auditing), store encrypted (secure storage with your keys), or no storage (nothing is kept after processing). This way you can always balance compliance needs with data privacy.

Your data is encrypted and stored on CLoD's infrastructure, with options to store raw, store encrypted, or not store at all. It's never used to train external models.

If you go over the Free Plan limits, your requests will pause until the next cycle, or you can upgrade to the Standard Plan at any time for higher limits and production-ready features. This way, you won't be billed unexpectedly, and you stay in control of when to scale up.

CLoD combines regex checks, PII detectors, context-aware AI (e.g. Llama Guard), and link filtering to identify, flag and allow, or block sensitive information before it leaves your system.

Yes, but it's minimal and won't affect user experience. Any real-time governance adds some overhead, and CLoD is built to minimize it while continuously improving performance.

For now, CLoD creates audit-ready logs directly viewable in the dashboard today for compliance review.

CLoD uses a three-step fallback system: (1) route to your chosen model if it's available, (2) if not, automatically reroute to the next best model, and (3) if a model doesn't respond, select another based on cost, latency, and capability, so requests always succeed.

FULL VERSION: CLoD protects against five key risks: Content Moderation (blocks toxic, harmful, or disallowed outputs), Data Leakage (stops sensitive or proprietary data from leaving), Policy Compliance (enforces your organization's rules automatically), Allowed Domains (filters unknown or suspicious links), and Prompt Defense (prevents jailbreaks or malicious instructions).

The CLoD dashboard gives you full visibility into your AI usage and performance. You can track the number of requests, threats detected, and overall violation rate, along with detailed charts for latency and time-to-first-token. These insights help teams monitor performance and audit compliance with ease.

A RAG (Retrieval-Augmented Generation) file is your organization's knowledge base. Upload it so you can (1) define policies CLoD enforces automatically, and (2) let your team chat with AI using your own trusted documents.

CLoD enforces your rules instantly through a layered approach. We combine regex pattern matching, prompt injection defenses, privacy protection tools (such as Microsoft Presidio), and industry-grade moderation models like OpenAI Moderation and Llama Guard. On top of this, CLoD uses a mix of retrieval-augmented generation (RAGs) and AI models to detect prompts that violate privacy or compliance standards. Together, these safeguards prevent risky data leaks and harmful outputs before they happen.

Yes. CLoD comes with built-in preset policies you can apply with one click. These range from strict lockdown for regulated industries (finance, healthcare, legal) to moderate filters for customer support chats, and even relaxed settings for internal testing. You can start with these presets and adjust as needed, or create your own custom guardrails from scratch.

Yes. CLoD supports adding your own OpenAI-compatible models through a pay-as-you-go setup. This means you can extend CLoD's guardrails, monitoring, and policy enforcement to the models you already use, without being locked into a single provider.

Yes. CLoD is designed to help organizations align with major governance and compliance frameworks out of the box, including AI RMF, EU AI Act, GDPR, HIPAA, SOC 2, CCPA, and ISO 42001. You can also customize policies to meet your own internal standards or regional requirements.

Yes. CLoD includes a Playground where you can experiment with models, policies, and guardrails in a safe test environment without affecting production systems. It's the easiest way to try things out before rolling them into your workflows.

Today, CLoD runs as a managed service on our infrastructure. We handle the reliability, monitoring, and updates so you don't have to. In the future, we plan to offer deployment options that let you run CLoD on your own infrastructure if needed.

CLoD currently supports two roles. Owner: full access and visibility, including managing policies, settings, and team members. Member: Members can use the API key, track usage history, and chat, but they can't change settings or remove others.

You can connect via direct API calls today, use the OpenAI SDK with minimal setup, and soon we'll release official SDKs and support for open standards like MCP (Model Context Protocol), a new way to plug tools together without extra coding.

CLoD currently supports 26 leading models from providers including OpenAI, Anthropic, Google, Fireworks, xAI, Together, Sambanova, Cerebras, and Groq. This includes well-known models like GPT-4.1, Claude Opus 4, Gemini 2.5, Grok 4, and multiple Llama variants. We're continuously adding new models, so you'll always have access to the latest and best-performing options.

Setup takes less than 5 minutes. You simply create an account, generate an API key, and drop it into your workflow. From there, you can start making governed AI requests right away after selecting a preset rulset or customizing your own.