# Indiseek — the full text > The complete content of the Indiseek beta page, in prose, for a language model that wants the whole thing in one document rather than a set of links. Every fact here is also on https://indiseek.com/coming-soon, in llms.txt and in the page's structured data. ## What Indiseek is Indiseek is an independent web search engine. It runs its own crawler, builds its own inverted index, and ranks with its own model: BM25 with field weights, fused with a link-graph authority score, plus freshness and spam scoring. It does not call another company's search API, and a results page makes no third-party request at all. Indiseek is a general web search engine. Anyone can search the open web with it, and any site on the web can be included in the index. Its position is deliberately narrow: most search boxes are a form in front of somebody else's index. Indiseek is the whole machine — the crawler, the storage engine, the ranking model, the API and the interface — engineered the way infrastructure is engineered, with deterministic builds, measured latency budgets, and an on-call engineer for every release. ## The facts, one line each - An independent search engine: its own crawler, its own inverted index, its own ranking model. It does not call another company's search API. - No ads and no data resale. Search terms are not sold, and no advertiser can buy a position in the results. - Zero third-party requests. A results page loads nothing from anyone else — no CDN, no webfont, no tracker, no analytics. - 2.4M+ pages indexed today, with a 118 ms median time to results, read from the live index rather than a marketing slide. - 23 languages and locales, including Devanagari to Latin transliteration, so a query typed in Hindi or Tamil finds a page written in English. - Private beta 15 December 2026, public beta Q1 2027. Free to use, built and operated from India by DiggajCoder. - Robots-respecting (RFC 9309) and IT Rules 2021 compliant, with a published grievance officer, a 24-hour acknowledgement SLA and a 15-day resolution SLA. ## What is inside the engine **Our own crawler.** WebSearchBot reads robots.txt to RFC 9309, discovers pages through sitemaps and RSS, and honours per-host politeness delays. It identifies itself with a public contact address. **Our own inverted index.** Postings, document frequencies and field-weighted term statistics are stored in our own engine — no third-party search API sits between a reader and a document. **Link graph and authority.** PageRank over outbound edges, combined with freshness and a spam-scoring model, so promoted junk cannot buy its way to the top. **Query understanding.** Spell correction, field-weighted BM25, reciprocal-rank fusion across signals, and Devanagari to Latin transliteration across 23 locales. **Privacy by construction.** No third-party requests, no fingerprinting on the results page, and no selling of search terms. The engine you query is the engine that answers. **Accountable and compliant.** A published grievance officer, 24-hour acknowledgement and 15-day resolution SLAs, and a takedown queue — as required by the IT Rules 2021. ## How it works — four stages, one pipeline 1. **Crawl.** Frontier scheduling, robots.txt compliance, per-host rate limits and sitemap discovery. The web is walked politely, at scale. 2. **Extract.** Boilerplate is stripped by text density, leaving title, body, links and the media a page declares — images and video included. 3. **Index.** Tokenised postings, document frequencies and field weights are written to our own inverted index, then linked into the page graph. 4. **Rank.** BM25 with field weights, fused with link-graph authority, freshness and spam scoring. Ranked results, served from our own origin. ## The privacy position Indiseek is funded as infrastructure, not by advertising or by selling search data. Search terms are not sold. The default is no cross-site tracking, no fingerprinting and no third-party analytics on the results page. A results page makes zero third-party requests — which means there is nothing on it that another company can see, sell or profile from. What the engine keeps is what it needs to crawl and rank the web: the pages, the links between them, and the signals inside the document. ## Language support The index and the interface support 23 locales, including Indian languages with Devanagari to Latin transliteration, so a query typed in Hindi or Tamil finds the page written in English. ## Roadmap - **November 2026 — private beta, by invitation.** A closed cohort of publishers, developers and researchers gets first access, with a direct line into the ranking team's backlog. - **15 December 2026 — the private beta opens, and waitlist invites go out.** Every waitlist entry is invited in batches, with a one-click form to submit a site for inclusion in the first index. - **Q1 2027 — public beta, free, for everyone.** Open access to the engine, the crawl dashboard and the operator tools that make a site owner's life easier. - **Q2 2027 — API, plugins and syndication.** A documented search API, a browser extension and an OpenSearch description so Indiseek can be the default engine anywhere. ## The questions people actually ask **What is Indiseek?** Indiseek is an independent web search engine. We run our own crawler, we build our own inverted index, and we rank with our own model (BM25 with field weights, fused with a link-graph authority score). We do not call another company's search API, and a result page makes no third-party request at all. **When does the beta open?** The invite-only private beta opens on 15 December 2026, and the free public beta follows in the first quarter of 2027. Everyone on the waitlist gets an invitation and a one-click site-submission form. **Will Indiseek show ads, or sell my data?** No. Indiseek is funded as infrastructure, not by advertising or by selling search data. Search terms are not sold, and the default is no cross-site tracking, no fingerprinting and no third-party analytics on the results page. **How is my site included in the index?** Our crawler follows robots.txt (RFC 9309), reads sitemaps and RSS feeds, honours per-host politeness delays, and identifies itself as WebSearchBot with a public contact address. You can also submit a URL directly for faster inclusion. **Which languages are supported?** The index and the interface support 23 locales, including Indian languages with Devanagari to Latin transliteration, so a query typed in Hindi or Tamil finds the page written in English. **Who builds Indiseek?** Indiseek is engineered by DiggajCoder, an enterprise software company. The whole stack is in-house: the crawler, the storage engine, the ranking model, the API, the web interface and the compliance desk. **Is Indiseek an alternative to Google Search?** Yes, with one honest caveat: an independent index is a young index. Indiseek does not call Google's API and does not borrow its ranking; it crawls and ranks the web itself, so its coverage starts small and grows every day. No advertiser can buy a position and there are no sponsored results. **Is Indiseek free?** The private beta and the free public beta are free to use, with no ads and no paid placement. A paid tier for the search API, the browser extension and syndication is on the roadmap for Q2 2027; the search engine itself stays free. **Does Indiseek log what I search for?** No. A results page makes zero third-party requests, there is no cross-site tracking and no fingerprinting, and search terms are not sold or shared. What the engine keeps is what it needs to crawl and rank the web: the pages, the links between them, and the signals in the document. **How is Indiseek different from DuckDuckGo, Brave Search or Kagi?** Every one of those engines is a different set of trade-offs. The difference here is that the whole stack is ours — the crawler, the storage engine, the ranking model, the API and the interface — and the business model is infrastructure rather than advertising or search-data resale. We do not call another company's search API, and there is no third-party request on the results page. **Can I get my site into the Indiseek index?** Yes. Our crawler discovers pages through sitemaps and RSS feeds and obeys robots.txt to RFC 9309, and you can also submit a URL directly for faster inclusion. There is no fee and no placement guarantee: a page earns its position on relevance, freshness and authority. **When will Indiseek be open to everyone?** The invite-only private beta opens on 15 December 2026 and the free public beta follows in the first quarter of 2027. The waitlist is the only queue: everybody on it is invited, and a one-click site-submission form comes with the invitation. ## Joining the beta The waitlist is the only queue. One email at launch, an invitation to the private beta, and a place in the roadmap — no drip campaign, and no reselling of your address. The address is stored once, in our own database, for one purpose: the launch email. No IP address, no user agent, no fingerprint, no resale. A reader who prefers a direct line can write to contact@indiseek.com and be added to the beta list by hand. ## Compliance and contact Indiseek is a general intermediary under the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021. The published obligations live at https://indiseek.com/rules, with the privacy policy at https://indiseek.com/privacy, the user agreement at https://indiseek.com/agreement, the grievance officer at https://indiseek.com/grievance and the takedown process at https://indiseek.com/takedown. The crawler policy, including how to submit a URL, is at https://indiseek.com/crawler. Contact: contact@indiseek.com. Engineered by DiggajCoder — https://diggajcoder.com.