NM Signals vs Reference: Features, Pricing & Which Is Better (2026)

A side-by-side comparison of NM Signals and Reference — features, pricing, and ideal use cases — to help you decide which AI tool fits your workflow.

NM Signals

Nyman Media

Freemium

Audits whether AI crawlers and assistants can actually read your website, then tracks how often they mention your brand.

AI Readiness Audit: Scores a public URL across 106 checks in six categories — crawlability, structured data, entity clarity, content structure, answerability and trust signals — for a readiness score out of 100.
Served-vs-Rendered Comparison: Measures how much of the browser-rendered page survives a fetch-only request, flagging JavaScript-dependent content that non-rendering AI crawlers never see.
AI Crawler Access Checks: Reports robots.txt, canonicals, redirects and status codes specifically for AI crawlers such as OAI-SearchBot, not just traditional search bots.
UX Review with Developer Brief: Runs a separate usability pass with visual layout analysis on paid plans and produces a copyable brief a developer can work straight from.
Saved Action Plans: Keeps an audit as a private baseline, lets you rank findings by priority, and records implementation progress against it.
Generated Fixes and Verification: Premium plans generate implementation guidance for a selected finding and verify the change against a fresh audit rather than trusting a checkbox.
AI Answer Snapshots: Asks the same five core questions weekly with three samples each, deciding by majority whether the brand is named, and charts the trend against tracked competitors.
Programmable Surface: A public REST API, CLI and MCP server let audits run inside CI/CD pipelines or be called directly by AI agents.

AI Search Readiness: Find out why an AI assistant summarizes a competitor's page instead of yours and fix the specific access or rendering issue behind it.
Pre-Launch QA: Audit a new marketing site before launch to catch blocked crawlers, missing markup and unreadable server-rendered content.
CI/CD Regression Guards: Call the REST API or CLI on every deploy so a rendering change that hides content from crawlers fails the build.
Brand Monitoring: Track weekly whether AI assistants name your brand in answers to the questions your buyers actually ask.
Agency Reporting: Produce white-label PDF audits and score comparisons for client sites on the Partner plan.
Content Restructuring: Use heading/body agreement and attribution checks to rewrite pages into retrievable, quotable sections.

Rahul Thennarasu

Free

Local, offline semantic search for your files and code — built for AI agents like Claude Code, with an MCP server.

Local Semantic Index: Runs a local embedding model over your files and code so nothing leaves your machine.
Live Auto-Reindexing: Updates the index as you save files, keeping search results current with the code you are actively writing.
Code-Aware Chunking: Uses tree-sitter to chunk on functions and syntactic units, so results are cited down to the exact code region.
Built-in MCP Server: Exposes /search, /explain, /find_similar and /check_doc_drift endpoints so Claude Code and other agents can query directly.
Cited Answers: Every result points back to a file and function, replacing generic AI advice with grounded, verifiable references.
Doc Drift Checks: /check_doc_drift flags places where documentation has fallen out of sync with the underlying code.
Cross-Codebase Similarity: /find_similar surfaces analogous implementations elsewhere in the codebase for reuse and refactoring.

Grounded Coding Q&A: Ask 'how did I implement rate limiting here' and get the actual function back, not a generic explanation.
Claude Code Context Injection: Use the MCP server so Claude Code pulls precise cited snippets instead of running expensive grep loops.
Refactoring Prep: Use /find_similar to locate analogous implementations across the repo before consolidating or standardizing them.
Doc-Code Alignment: Run /check_doc_drift to catch documentation that no longer matches the code it describes.
Offline / Air-Gapped Work: Semantic code search on machines that can't upload source to cloud embedding services.
Cross-Project Recall: Reindex multiple repos locally to find prior solutions you already wrote in another project.