Independent technical audit

carbonforge.dev

Overall AI readiness score

A combined result across all audit signals.

Excellent

AI readiness audit: 10/28/2026Public technical data

Overall AI readiness assessment

How ready carbonforge.dev is for AI systems

86/100

Carbon Forge is a legitimate source-control platform for humans and AI agents. Its llms.txt is accessible, well-formed, and consistent with the homepage. No policy issues found; approved for the directory.

Audit context

llms.txt is accessible and contains 1,247 tokens. No accessible llms-full.txt was detected, so deeper context still has to be assembled from regular pages. No separate ai.txt policy was detected; it is optional, but can clarify training, retrieval, and attribution preferences.

robots.txt is available. 11 of 11 tracked AI bots are not blocked. Declared sitemaps: 1. The homepage exposes Schema.org types Organization, WebSite, SoftwareApplication; 9 OpenGraph tags were found and markup completeness is 90%.

The mobile Lighthouse profile adds Performance 100/100, Accessibility 100/100, Best Practices 100/100, SEO 100/100, and experimental Agentic Browsing 100/100. These signals have a limited weight: they complement rather than replace llms.txt, robots.txt, and structured-data checks.

Confirmed strengths

Accessible llms.txt
1 sitemap declaration
Schema.org: Organization, WebSite
Complete social metadata
Strong Google agentic signals

Priority improvements

1Add llms-full.txt for richer first-party context.

llms.txt tokens1,247

llms-full.txt tokens—

ai.txt

sitemap.xml

Google Lighthouse technical profile

A mobile Lighthouse measurement. Google’s experimental Agentic Browsing category is explained separately and does not replace the broader llmsmap AI-readiness score.

Mobile · Lighthouse

100

Performance

100

Accessibility

100

Best Practices

100

Technical SEO

100

Agentic Browsing

What these results mean

The mobile page renders efficiently; its largest visible content block appeared in 0.9 s.

Accessibility scored 100/100, Best Practices 100/100, and technical SEO 100/100. The experimental Agentic Browsing category scored 100/100. It measures signals Google currently tests for software agents and is shown separately from the llmsmap AI-readiness score.

FCP0.9 s

First content

LCP0.9 s

Main content

CLS0

Layout stability

TBT0 ms

Blocking time

SI1.1 s

Visual speed

Metric glossary

FCP · First content: When the first text or image appeared on screen.
LCP · Main content: When the largest visible element in the first viewport rendered.
CLS · Layout stability: How much content shifted unexpectedly while loading; lower is better.
TBT · Blocking time: How long the browser main thread could not respond quickly to input.
SI · Visual speed: How quickly the visible viewport filled with content.

10/28/2026Lighthouse 13.5.0Mobile profile

AI readiness checks

Machine-readable files, crawler policy, discovery, and homepage markup.

llms.txt

File found and accessible

https://carbonforge.dev/llms.txt

llms-full.txt

Full version was not found

ai.txt

ai.txt file was not found

robots.txt

File found

https://carbonforge.dev/robots.txt

Sitemap in robots.txt1

1 sitemap found

Schema.org (JSON-LD)

Types: Organization, WebSite, SoftwareApplication

OpenGraph90%

9 OG tags found

AI bot access

Based on robots.txt analysis

GPTBotAllowed

OAI-SearchBotAllowed

ChatGPT-UserAllowed

Google-ExtendedAllowed

ClaudeBotAllowed

Claude-SearchBotAllowed

Claude-UserAllowed

BytespiderAllowed

CCBotAllowed

PerplexityBotAllowed

Perplexity-UserAllowed

Sitemaps

Declared discovery routes for crawlers and agents.

https://carbonforge.dev/sitemap.xml

OpenGraph tags

Social preview metadata found on the homepage.

Markup completeness: 90%

og:typewebsite

og:site_nameCarbon Forge

og:titleCarbon Forge — the record your agents read and write

og:descriptionCode, knowledge and the reason for every change in one sealed record that agents query with a call. Git is rendered from it.

og:urlhttps://carbonforge.dev/

og:localeen_US

og:imagehttps://carbonforge.dev/_site/og/home.png

og:image:width1200

og:image:height630

Schema.org markup

Structured entities and properties exposed on the homepage.

Types found: 3 · Properties: 6

OrganizationWebSiteSoftwareApplication

name100monkeys AI, Inc.

urlhttps://100monkeys.ai

logohttps://100monkeys.ai/aegis-logo.png

sameAshttps://github.com/100monkeys-ai, https://www.linkedin.com/company/112423436/, https://x.com/100monkeysai, https://hashnode.com/@100monkeys-ai, https://medium.com/@jeshua_46811

publisher100monkeys AI, Inc.

descriptionThe sealed record of code and knowledge that agents read and write with a call. Git is rendered from it.

# Carbon Forge

> Carbon Forge keeps code, files and prose in one record on the server, and people and agents read and write it through the same tools. Agents reach it over MCP; git is rendered from the record for the tools that still speak git. Public workspaces are free, and every account includes one free private workspace.

Carbon Forge is built by 100monkeys AI, Inc. on Carbon, the source-control model its docs describe. An agent opens a lane instead of cloning: a lane is a set of patches over the trunk, and every read on it composes the trunk as it stands now with those patches. Carbon Forge never merges text by itself; a patch that no longer applies is reported as a named conflict. A landing is one changeset: every page, atom, code tree and file it names is written in one transaction, or nothing is. Every revision and every changeset is hash-chained, and a revert is a new landing, so history only moves forward. Blame, grep, diff and history run on the server.

The docs at https://carbonforge.dev/docs/p/ are rendered in the browser by the app, so this file links each page rather than reproducing it in an llms-full.txt.

## Site

- [Carbon Forge](https://carbonforge.dev/): the record your agents read and write
- [Pricing](https://carbonforge.dev/pricing): Free, Pro, Team and Enterprise plans
- [Security](https://carbonforge.dev/security): a history you can verify
- [Specification](https://carbonforge.dev/specification): the Carbon SCM specification
- [Carbon Forge and AEGIS](https://carbonforge.dev/aegis)
- [Comparisons](https://carbonforge.dev/compare): side by side, every row cited
- [Sign in](https://carbonforge.dev/login): a one-time link by email, no passwords
- [FAQ](https://carbonforge.dev/faq): short answers on agents sharing one codebase, shared memory, auditing agent changes, MCP for a knowledge base, and git for teams of agents
- [Guides](https://carbonforge.dev/guides): step-by-step guides built from the docs
- [Land a code change and its documentation in one changeset](https://carbonforge.dev/guides/code-and-docs-in-one-changeset): change code and its page on one lane and land both as one changeset
- [Cut a release and push it to GitHub](https://carbonforge.dev/guides/cut-a-release-and-push-to-github): name a release with a cut, read its notes, render it as git and push fast-forward only
- [Catch stale docs with governed seams](https://carbonforge.dev/guides/governed-seams-drift-report): declare which pages govern which code and read a drift report before a landing
- [Connect Claude Code to Carbon Forge over MCP](https://carbonforge.dev/guides/claude-code-mcp): attach Claude Code with a scoped token so agents read, write and cite one record
- [Conecta Claude Code a Carbon Forge por MCP](https://carbonforge.dev/guides/es/claude-code-mcp) (es): the same guide in Spanish
- [Conecte o Claude Code ao Carbon Forge via MCP](https://carbonforge.dev/guides/pt/claude-code-mcp) (pt): the same guide in Portuguese
- [Claude Code を MCP で Carbon Forge に接続する](https://carbonforge.dev/guides/ja/claude-code-mcp) (ja): the same guide in Japanese
- [Connect Cursor, Codex and any MCP client](https://carbonforge.dev/guides/cursor-codex-mcp): one token per client and the exact configuration for each
- [Agent sign-up skill](https://carbonforge.dev/skill): a Markdown skill an agent follows to create an instance for its user

## Docs

- [Welcome to Carbon](https://carbonforge.dev/docs/p/home): the docs home
- [Overview](https://carbonforge.dev/docs/p/overview): what Carbon Forge is and is not
- [Source control on the substrate](https://carbonforge.dev/docs/p/features/source-control-on-the-substrate): the problem it solves and the design
- [Working in lanes](https://carbonforge.dev/docs/p/guides/lanes)
- [Code and files on the substrate](https://carbonforge.dev/docs/p/guides/code-and-files)
- [Source control: how it is built and why](https://carbonforge.dev/docs/p/architecture/source-control)
- [A shared cortex for LLMs](https://carbonforge.dev/docs/p/features/shared-cortex-for-llms): MCP as a primary product surface
- [cortex.ground](https://carbonforge.dev/docs/p/guides/cortex-ground): how an agent orients itself
- [MCP tools catalog](https://carbonforge.dev/docs/p/api/mcp-tools-catalog)
- [Billing and plans](https://carbonforge.dev/docs/p/guides/billing-and-plans)

## Comparisons

- [Carbon Forge vs GitHub](https://carbonforge.dev/compare/github)
- [Carbon Forge vs GitLab](https://carbonforge.dev/compare/gitlab)
- [Carbon Forge vs Notion](https://carbonforge.dev/compare/notion)
- [Carbon Forge vs Obsidian](https://carbonforge.dev/compare/obsidian)
- [Carbon Forge and AEGIS vs CI/CD platforms](https://carbonforge.dev/compare/ci-cd)
- [Carbon Forge and Anthropic's Scion](https://carbonforge.dev/compare/anthropic)

## Company

- [100monkeys.ai](https://100monkeys.ai/): the company that builds Carbon Forge, AEGIS and Your Zaru
- [GitHub](https://github.com/100monkeys-ai)
- [X](https://x.com/100monkeysai)

Share or continue the analysis

Share this audit