🛠️ What Is This Toolkit? (Direct Answer)
This is a stack of 10 free, open-source repositories on GitHub — built and used inside a live SEO/GEO practice — that automate the parts of technical SEO, AI-search visibility, and content production that used to be manual or locked behind paid platforms. The stack covers AI citation tracking, AI-crawler robots.txt auditing, full-site SEO audits, bulk structured data validation, a config-driven automation framework, a multi-stage content pipeline, Google Search Console sitemap submission, and a data analysis agent — plus two general-purpose utilities. Most of the AI-powered tools use a bring-your-own-key (BYOK) model: the code is free, and you pay your LLM provider only for what you run.
All 10 repositories are public on GitHub under the indexcraft organization. Repo names, descriptions, and links are listed in full in the table below.Most SEO tooling still ships as a subscription dashboard: pay monthly, get a fixed set of reports, and hope the vendor adds the feature you actually need before your renewal comes up. That model works fine for backlink and keyword-volume data, where the value is in a proprietary crawled index nobody can replicate solo. It works far worse for the newer problems in this field — tracking AI citations, auditing AI-crawler access, validating structured data at scale, orchestrating a content pipeline — where the logic is closer to a well-written script than a proprietary database.
Over the last year I moved a growing share of my own SEO/GEO operations onto a set of scripts and small agents I could actually read, modify, and run on my own schedule. This guide walks through all 10 of those repositories, what each one is for, and how they combine into a single working stack — not as a sales pitch, but as a practitioner's map of what's actually useful, what's still rough around the edges, and where a paid platform is still the better call. For the broader technical SEO foundation these tools sit on top of, see the Technical SEO Guide and the Complete SEO Guide 2026.
- All 10 tools are free and open source; several follow a bring-your-own-API-key model, so the only ongoing cost is your own LLM usage.
- They split into five functional groups: AI search visibility, SEO audit & automation, content production, Search Console tooling, and data analysis, plus two bonus utilities.
- This stack is built to sit alongside a commercial platform, not replace one — it fills GEO and automation gaps most SaaS tools haven't caught up to yet.
- Being open source is the actual security advantage for anything touching a Google Search Console service account: you can read exactly what the code does before granting access.
- Start with the tool that removes a task you already do by hand every week — usually the structured data auditor or the GSC sitemap submitter.
- None of this requires deep engineering skill — basic command-line comfort (clone, install, set an environment variable) is enough to run every repo here.
1. Why Build an Open-Source SEO/GEO Stack
Three specific frustrations pushed this stack into existence. First, GEO monitoring — tracking whether ChatGPT, Claude, Perplexity, or Gemini actually cite a brand in their answers — is a category most commercial rank trackers still bolt on as an afterthought, if they offer it at all. Second, AI-crawler access auditing needs a constantly updated list of named bots (GPTBot, ClaudeBot, PerplexityBot, and dozens more), which changes faster than most robots.txt checkers get updated. Third, a lot of routine SEO operations work — resubmitting sitemaps across properties, validating JSON-LD across hundreds of URLs, scheduling recurring checks — is exactly the kind of repetitive task a small script handles better than a person clicking through a dashboard.
None of that requires a large, continuously crawled web index — the thing that actually justifies platforms like Semrush or Ahrefs charging what they do. It requires clear logic, a scheduler, and in a few cases, an LLM call. That's a solo-buildable problem, and it's why this ended up as a set of open repositories rather than another subscription.
None of these tools started as a plan to "build an SEO toolkit." Each one started as a specific Tuesday-afternoon annoyance — resubmitting the same sitemap across four Search Console properties by hand, or realizing a client's FAQPage schema had been silently broken for three weeks because nobody was checking it at scale.
The pattern that emerged: whenever a task repeated more than three or four times in a month, it became a candidate for a script. The content-production-pipeline repo came last, once enough of the smaller pieces existed that stitching them into a five-stage workflow made sense rather than being a rebuild from scratch.
The honest lesson: build the smallest tool first, use it enough to find its rough edges, and only then decide whether it's worth generalizing. Several of these repos are still narrow by design — that's a feature, not an unfinished feature.
2. The 10 Tools at a Glance
Here's the full stack in one table, grouped loosely by what problem each tool solves. Every tool name links directly to its GitHub repository.
| Tool | What It Solves | Category | Cost Model |
|---|---|---|---|
| ai-citation-tracker | Tracks whether a brand is cited by ChatGPT, Claude, Perplexity, and Gemini, with position and competitor share-of-voice | AI Search Visibility | Free + BYOK |
| ai-crawler-access-toolkit | Audits robots.txt against 24 named AI crawlers; validates and generates llms.txt | AI Search Visibility | Free |
| ai-seo-audit-agent | Full-site technical, content, and GEO/AI-crawler audit in one command, with optional LLM executive summaries | SEO Audit & Automation | Free + optional BYOK |
| structured-data-bulk-auditor | Bulk-validates JSON-LD across a sitemap or URL list against schema.org rules | SEO Audit & Automation | Free |
| seo-ops | Config-driven scheduling, webhook server, and multi-channel notifications for recurring SEO tasks | SEO Audit & Automation | Free |
| content-production-pipeline | Five-stage virtual content team producing a publish-ready article with FAQ schema and JSON-LD | Content Production | Free + BYOK |
| sitemap-submitter-gsc | Auto-submits multiple sitemaps across multiple Search Console properties via a service account | Search Console & Indexing | Free |
| data-analyst-agent | Queries, summarizes, and visualizes structured datasets, deterministically or via an optional LLM layer | Data Analysis | Free + optional BYOK |
| google-100-results | Browser extension showing 20–100 Google results per page, manually or automatically | Bonus Utility | Free |
| aarav-video-splitter | Splits any video into equal parts losslessly via a simple GUI, standalone Windows exe | Bonus Utility | Free |
3. AI Search Visibility & GEO Tools
These two repos address the newest and least standardized part of the stack: knowing whether AI systems can find and cite you at all. For the underlying concept, see the AI Citation Rate Guide and the guide to optimising for Perplexity, ChatGPT, and Gemini.
ai-citation-tracker
ai-citation-tracker queries ChatGPT, Claude, Perplexity, and Gemini with a defined set of prompts relevant to your brand or client, then logs whether your brand gets cited, where it appears in the response, and how that compares to competitors mentioned in the same answer. The output is a citation rate, a rough position signal, and a share-of-voice comparison — the closest GEO equivalent to a traditional keyword rank tracker, except the thing being measured is presence inside a generated answer rather than a position in a results list.
This is the tool I'd point to first for anyone asking "does AI search even matter for my site yet" — running it for a month against your actual brand and competitor set turns that question into a number instead of a guess.
ai-crawler-access-toolkit
ai-crawler-access-toolkit checks your robots.txt against 24 named AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others — and separates bots used for model training from bots used for live search-answer retrieval. That distinction matters: blocking a training-only crawler is a legitimate content-licensing choice, while accidentally blocking a search-retrieval crawler can quietly remove you from AI-generated answers without moving your Google rankings at all. The toolkit also includes an llms.txt validator and generator, for sites experimenting with that emerging standard. For the full policy discussion, see the robots.txt for AI Crawlers Guide and the llms.txt Guide.
4. SEO Audit & Automation Tools
Three repos handle the recurring, checklist-shaped work of technical SEO — the kind of task that's tedious by hand and a natural fit for a script running on a schedule.
ai-seo-audit-agent
ai-seo-audit-agent runs a full-site audit — technical checks, content-quality checks, and GEO/AI-crawler-visibility checks — in a single command. An optional LLM layer, using your own OpenAI or Anthropic key, turns the raw findings into a consultant-style executive summary, which is genuinely useful when the audience for the audit isn't technical. This overlaps with the manual workflow in the SEO Audit Guide — the agent automates the repeatable checks from that process; the judgment calls on prioritization still belong to a human.
structured-data-bulk-auditor
structured-data-bulk-auditor validates JSON-LD structured data across an entire sitemap or URL list against schema.org rules, flagging missing required fields and broken FAQPage, BreadcrumbList, or Product structures before they cost you rich-result eligibility. This is the tool I'd recommend running immediately after any CMS migration or template change — schema errors introduced by a template update tend to be silent and site-wide, which is exactly the failure mode a bulk auditor is built to catch. See the Schema Markup Guide 2026 for the full implementation reference this tool checks against.
seo-ops
seo-ops is the scheduling backbone that ties recurring checks together: a config-driven framework using APScheduler for timing, a Flask webhook server for triggering runs from tools like n8n, YAML-based configuration, and multi-channel notifiers so a failed check actually reaches you instead of sitting unread in a dashboard. It's less a single-purpose tool and more the glue that turns the other repos into something that runs itself on a schedule rather than something you remember to run manually.
5. Content Production Pipeline
content-production-pipeline is the most involved tool in the stack: a five-stage virtual content team orchestrated via an LLM (bring your own OpenAI or Anthropic key), producing a publish-ready article complete with front matter, FAQ schema, and JSON-LD.
Gathers source material and context for the target topic before any drafting begins.
Structures the research into a logical heading hierarchy and section plan.
Drafts full content against the outline, stage by stage.
Reviews the draft for coherence, accuracy of structure, and consistency before it moves downstream.
Adds front matter, FAQPage and Article JSON-LD, and a final technical-SEO pass.
The point of splitting this into five discrete stages rather than one long prompt is the same reason a human editorial team doesn't hand a single writer the entire process end to end: each stage has a narrower job, which makes the output more consistent and each step easier to inspect or override individually. It's still a starting draft, not a final one — see the Limitations section below on what human review this doesn't replace.
6. Search Console & Indexing Tool
sitemap-submitter-gsc
sitemap-submitter-gsc automatically submits multiple sitemaps to Google Search Console across multiple properties using a service account, removing the manual, one-property-at-a-time submission flow inside the GSC interface. For agencies or in-house teams managing several properties — regional domains, subdomains, or a portfolio of client sites — this turns a monthly chore into a scheduled job. Pair it with seo-ops for scheduling, and see the Google Search Console Guide for the manual workflow it's replacing.
7. Data Analysis Agent
data-analyst-agent
data-analyst-agent queries, summarizes, and visualizes structured datasets — exported GSC data, crawl exports, analytics pulls — and works deterministically without any API key. An optional bring-your-own-LLM layer adds natural-language insight generation on top, for cases where you want a plain-English summary of what a dataset shows rather than reading a chart yourself. The deterministic base mode matters: you can run this on sensitive client data without ever sending it to a third-party LLM provider, and only opt into the LLM layer when that's actually acceptable for the dataset in question.
8. Bonus Utilities
Two repos in the org aren't strictly SEO tools but earn their place in the same toolbox for the practical, adjacent problems they solve.
google-100-results
google-100-results is a Chrome, Edge, and Firefox extension that shows 20 to 100 Google search results on a single page, manually or automatically, instead of paginating 10 at a time. For manual SERP research — competitor scanning, feature spotting, or just seeing a fuller results picture without repeated pagination — it removes a genuinely annoying piece of friction.
aarav-video-splitter
aarav-video-splitter splits any video into equal parts losslessly — no re-encoding, no quality loss — through a simple GUI, packaged as a standalone Windows executable. It's a general-purpose utility rather than an SEO tool specifically, useful for anyone chopping long-form video into clips for short-form or platform-specific distribution as part of a broader content workflow.
9. How These Tools Fit Into One Workflow
Individually, each repo solves one problem. Run together on a schedule, they form a loop that mirrors how a small in-house SEO/GEO operation actually runs week to week.
structured-data-bulk-auditor and ai-crawler-access-toolkit run on a schedule via seo-ops, flagging any schema breakage or accidental AI-crawler blocking introduced by that week's deploys.
sitemap-submitter-gsc resubmits updated sitemaps across all managed properties after new content ships, keeping Search Console's picture of the site current.
ai-seo-audit-agent runs a complete pass, and ai-citation-tracker logs that month's AI citation rate against the same competitor set, so both technical health and AI visibility get tracked as trend lines rather than one-off snapshots.
content-production-pipeline produces a first draft for a planned topic; a human editor reviews and revises before publishing, and the new URL is added to the next sitemap submission cycle.
data-analyst-agent pulls the accumulated GSC and audit exports into a summary for stakeholder reporting, without needing GSC data to leave your own machine unless you choose to enable the LLM layer.
10. Getting Started: Installation & Keys
None of these tools ship as installers or SaaS dashboards — each is a repository you clone and run yourself, which is precisely what keeps the cost model at zero beyond your own API usage. The general pattern is the same across the stack:
1. Clone the repo: git clone https://github.com/indexcraft/<repo-name> 2. Read the README: each repo documents its own exact requirements 3. Install dependencies: Python repos typically use pip + a virtual environment 4. Set API keys as environment variables — never hard-code them in the script 5. Run the entry command documented in that repo's README
For the LLM-powered tools — ai-citation-tracker, content-production-pipeline, and the optional layers in ai-seo-audit-agent and data-analyst-agent — the bring-your-own-key model means you connect an OpenAI or Anthropic API key and pay that provider's standard token pricing directly. There's no markup and no subscription on top. Budget accordingly if you're running the content pipeline or citation tracker at scale across many prompts or many articles.
11. Limitations — What This Stack Doesn't Replace
Being upfront about the gaps here matters more than the pitch. This toolkit does not replicate the enormous, continuously crawled backlink and keyword-volume indexes that justify what platforms like Semrush and Ahrefs charge — building that from scratch as a side project isn't realistic, and this stack doesn't try to. Most practitioners running this stack still keep a commercial platform for that category of data.
✅ Honest Limitations to Plan Around
- These are solo/small-team-maintained open-source tools — no enterprise SLA, no dedicated support line
- Content pipeline output is a strong first draft, not a publish-ready final piece — human editorial review is still required
- AI citation tracking reflects a sampled set of prompts, not a comprehensive real-time index of every possible query
- Requires basic command-line comfort — cloning a repo, setting environment variables, running a script
- LLM-powered tools carry usage cost proportional to how heavily you run them, even with no license fee
- Not a replacement for a full backlink or keyword-volume index — pair with a commercial platform for that data
12. Frequently Asked Questions About Open-Source SEO & GEO Tools
What does GEO tooling actually do differently from traditional SEO tools?
Traditional SEO tools measure rankings in a list of blue links. GEO (Generative Engine Optimization) tooling measures something structurally different: whether an AI system — ChatGPT, Claude, Perplexity, Gemini — cites your brand at all inside a generated answer, and how you compare to competitors who got cited in the same response. There is no fixed position 1 through 10 to track; the unit being measured is citation presence and share-of-voice, not rank position.
Tools like an AI citation tracker query these systems directly and log whether and how a brand shows up.
Are these open-source SEO tools really free to use?
The code itself is free and open source under their published licenses. The cost that does exist is usage-based: several of these tools follow a bring-your-own-key (BYOK) model, meaning you connect your own OpenAI or Anthropic API key and pay that provider directly for the tokens you consume.
There is no subscription fee to IndexCraft or any markup on top of the LLM provider's own pricing — you only pay for what you actually run.
Do I need to know how to code to run these tools?
Basic command-line comfort helps — cloning a repository, creating a virtual environment, and setting environment variables for API keys. None of these are packaged as one-click SaaS dashboards.
If you can follow a README and are comfortable running a script from a terminal, you can operate every tool in this stack.
Can these tools replace a paid platform like Semrush or Ahrefs?
No, and that isn't really the goal. Platforms like Semrush and Ahrefs maintain enormous, continuously updated backlink and keyword indexes that would be impractical to replicate as a side project.
This toolkit fills a different, narrower gap: AI citation monitoring, AI crawler access auditing, structured data validation, and workflow automation — categories where most commercial platforms are either absent, expensive, or several steps behind. Many practitioners run this stack alongside a commercial platform, not instead of one.
Which tool should I start with if I'm new to SEO automation?
Start with whichever tool removes a task you currently do by hand every week. For most sites, that's the structured data bulk auditor (catching schema errors before they cost you rich results) or the sitemap submitter for Google Search Console (removing manual sitemap resubmission across properties).
Both have a narrow, well-defined job, which makes them the easiest starting point before moving on to multi-stage tools like the content production pipeline.
Is it safe to connect a Google Search Console service account to an open-source script?
Treat any service account credential the way you'd treat a password: read the script before running it, scope the service account to only the Search Console properties it needs, and never commit the credential JSON file to a public repository.
Because these are open-source tools, you can actually read exactly what the code does with the credential before granting access — which is more transparency than most closed-source SaaS integrations offer.
How is an AI crawler access audit different from a normal robots.txt check?
A normal robots.txt validator checks syntax against Googlebot and Bingbot rules. An AI crawler access audit checks your directives against a much longer list of named AI agents — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others — and distinguishes between bots used for model training versus bots used for live search-answer retrieval.
Blocking the wrong one can silently remove you from AI-generated answers while leaving traditional Google rankings untouched, which is why this is worth auditing separately.
What does a "bring your own LLM" content pipeline mean in practice?
It means the pipeline orchestrates the workflow — research, outline, draft, edit, SEO pass — but you supply the model that powers each stage via your own API key.
This keeps the tool provider-agnostic: swap between OpenAI and Anthropic models without switching tools, and keep full visibility into what each stage actually sends to the model, since the orchestration logic is open source rather than hidden behind a SaaS API.
How This Toolkit Connects to a Broader SEO & GEO Strategy
These tools automate execution — they don't replace the strategic layer underneath. The guides below cover the disciplines this toolkit sits on top of.
The metric the ai-citation-tracker repo measures — how to define it, benchmark it, and act on it across ChatGPT, Claude, Perplexity, and Gemini.
Read the full guide →The policy reference behind the ai-crawler-access-toolkit repo — every major AI crawler, training vs retrieval access, and directive examples.
Read the full guide →The rules the structured-data-bulk-auditor checks against, with implementation templates for every major schema type.
Read the full guide →The manual workflow the sitemap-submitter-gsc repo automates — every report, metric, and setup step covered in full.
Read the full guide →The full manual audit methodology behind ai-seo-audit-agent — technical health, content gaps, and AI visibility checks explained step by step.
Read the full guide →The companion guide on using Claude specifically as the LLM layer behind BYOK tools like content-production-pipeline and ai-citation-tracker.
Read the full guide →📚 Where the Links and Claims Come From
| Source | Key Relevance |
|---|---|
| github.com/indexcraft | The organization hosting all 10 repositories covered in this guide, including each repo's own README and license. |
| Google Search Central — Overview of Google Crawlers | Reference documentation for how Googlebot and related crawlers identify themselves, relevant to the AI-crawler auditing logic in this stack. |
| Google Search Console API Documentation | The official API reference behind service-account-based automation, relevant to sitemap-submitter-gsc. |
| Schema.org — Structured Data Vocabulary | The vocabulary and rule set structured-data-bulk-auditor validates JSON-LD markup against. |
| Sharma, R. (2026) — IndexCraft internal workflow notes | Practitioner experience from building and running this stack across IndexCraft's own SEO/GEO operations. IndexCraft internal research (data on file). |