NOAN

About NOAN

NOAN is a fact layer platform that turns a company's approved business information-pricing, positioning, policies, customer data-into a verified, versioned source of truth. It serves these facts to both people and AI agents through an API and MCP server, aiming to keep all tools and team members aligned on the same company decisions. The platform also includes a social-style Activity feed that logs every read and write across the system.

Review

NOAN approaches the problem of AI consistency by separating "approved facts" from the general mass of company documents. Instead of letting each AI agent interpret scattered files differently, teams define and verify specific blocks of truth that every connected model or app can pull from. The tool is built for organizations that already run multiple AI agents and need a single reference point, though it also ships with its own Slack agent called Verity and an open-source agent pack.

Key Features

• Fact blocks with versioning and permissions: Teams create discrete, verified facts (e.g., pricing, brand tone, refund policy). Each fact is versioned, and changes can require approval similar to pull requests on GitHub. Fact-level permissions control who can edit or verify.
• API and MCP server access: The fact layer is available through a REST API and a hosted MCP server. Developers connect tools like Claude, Cursor, or Codex in about a minute without an API key. The company's own website runs on NOAN, reading facts directly instead of using a traditional CMS.
• Activity feed: A timeline-style stream shows every read, write, and task performed by people and agents across the API. This lets teams trace any agent's answer back to the specific fact it used.
• Verity Slack agent: Verity answers questions from verified facts, drafts new facts when a user types "fact:", flags statements that contradict existing facts, and can manage tasks, contacts, and notes from Slack channels.
• Agent pack (MIT licensed): The makers publish the agents they use internally-for customer support, fact alignment, market research, weekly reports, and newsletters-on GitHub. Users bring their own model keys and run agents on their own infrastructure with no per-task metering.

Pricing and Value

NOAN offers a free sign-up tier. A paid Starter plan exists, and during the launch period a 50% discount code (PRODUCTHUNT50) applies for the first three months. Exact pricing figures for the Starter plan or any additional tiers are not listed on the current page. The agent pack is free and MIT-licensed, with users supplying their own LLM API keys and compute.

Pros

• Decouples verified business facts from raw documentation, which reduces conflicting answers across different AI tools.
• MCP server setup takes roughly a minute and doesn't require an API key for supported platforms.
• Activity feed creates an auditable trail of every fact read and write by both humans and agents.
• Agent pack ships under MIT license, letting teams inspect, modify, and self-host the automation code.
• Slack integration allows non-technical team members to interact with and contribute to the fact layer without leaving their chat environment.

Cons

• Teams without existing AI agent workflows may find the fact layer abstraction unnecessary compared to a shared document or wiki.
• Pricing details beyond the free tier remain undefined on the page, making cost projection difficult for larger deployments.
• Reliance on Slack for the Verity agent means teams using other chat platforms cannot access that interaction mode natively.

NOAN fits teams that already deploy multiple AI agents or assistants and struggle with each one producing slightly different answers from the same company material. It's less suited for organizations still working primarily with human-facing wikis and document folders, where the overhead of maintaining a separate fact graph may outweigh the consistency gains. The open-source agent pack and MCP integration point toward a developer-friendly setup, while the Slack agent and no-code app make the system accessible to non-technical staff who need to verify or query facts.

Open 'NOAN' Website