KolBo Search: an engine of our own — not a filtered feed

Somewhere in your community's network infrastructure, a benevolent lie is probably being told right now. A computer asks for www.google.com, and the network answers with a quiet substitution — a DNS record pointing the name at forcesafesearch.google.com, a stricter Google wearing the regular one's address. This is not a hack. It is Google's official, documented instruction for schools and families who need safe results guaranteed: rewrite the name system so the engine you don't quite trust resolves to a version of itself you trust slightly more. Bing publishes the same recipe — map www.bing.com to strict.bing.com. The mainstream's best answer to "make search safe" is, verbatim, impersonate us to ourselves.

Sit with how strange that is, because the entire architecture of kosher search — until now — lives inside that strangeness. And then read the one sentence on the kolbo.life homepage that steps outside it: KolBo Search is "a proprietary search engine, not a filtered feed."

Why search is the hardest surface on any device

A kosher device can omit social media, video apps, and news feeds — but the moment it carries a search box, the entire internet stands one query away. And unlike the feeds, search cannot simply be deleted, because it is where the legitimate life of a connected device happens: directions, phone numbers, the government form, the refund portal, the medical question at 11 p.m., the gemach's address. Search is simultaneously the most indispensable surface and the least predictable one — which is why every communal standard treats it as the hardest problem, and why the solutions built around it deserve a respectful, precise anatomy.

The workaround stack, layer by layer

What the community actually runs on today is scaffolding — locks stacked on an engine that was never its own. Each layer is real engineering; together they tell one story.

Layer one: the settings. Google's SafeSearch offers three states, and their names matter. "Filter" blocks all explicit results — it is applied automatically when Google detects a user may be under 18. "Blur" merely blurs explicit images while explicit text and links can still appear — and Blur, not Filter, is the default for adults. Bing's ladder is looser still: its default, "Moderate," screens images and video "but not text." Read those defaults again as a parent: out of the box, the mainstream engines' posture is partially covered. And Google states the boundary of the whole mechanism plainly: "SafeSearch only works on Google Search" — the pages behind the results are someone else's business.

Layer two: the locks on the settings. Because a setting a user can flip is a suggestion, the ecosystem's next floor is enforcement: the CNAME substitutions above, family-DNS services like CleanBrowsing and OpenDNS that auto-enforce them network-wide — and then, because a lock is only as good as its walls, the same providers instruct you to block every alternative engine, DoH resolver, and VPN that could walk around it. The mainstream's own architecture concedes the point this pillar rests on: the engine is the unsafe object; everything else is containment.

Layer three: the human screeners. The frum world, characteristically, went further than anyone: Israel's NetFree runs human review of images around the clock — staff manually approving what Google's results are allowed to show — and providers like Techloq check words and pictures before a page serves. This is the workaround stack at its most heroic: a community paying people to stand between its families and a results page, in real time, forever.

Layer four: the resignation. And still the trust never arrived. Google's own troubleshooting page concedes that "no filter is 100% accurate." The community's coffee-room threads say the quiet part — one YWN commenter, representative of hundreds: the protected browser "isn't perfect... you can still access things like google image search although it's forced safesearch." Everyone uses the stack. Nobody believes in it. That is the precise emotional state of kosher search in the overlay era.

“Twenty years of locks, screeners, and DNS tricks — all built on one shared assumption: the engine itself could never be ours.”

kolbo.life

The kid-safe aisle, and why a family can't live in it

One more shelf deserves its anatomy before the alternative, because parents find it first: the children's search engines. Kiddle presents a cartoon-robot page whose results ride Google Custom Search — its own about page describes the tiering: the first results hand-picked by editors from kids' sites, the next few from trusted simple-language sites, everything past that passed through Google's safety machinery. KidzSearch — a two-decade veteran and American Library Association member — wraps the same idea in a school portal of videos, games, and facts, with locked strict settings that made it a classroom standard. On their own terms, these are honorable products. Swisscows, from Switzerland, takes the values a level deeper — explicit content "neither indexed nor displayed," with no off-switch offered — but remains a generic engine with a generic index, built for no particular community's life. And DuckDuckGo, often recommended in privacy circles, offers a safety toggle with no lockable parental control at all — its own enforcement guidance is, once again, a DNS substitution.

Here is why none of it answers the frum family's question: a household cannot run on a children's sandbox. The kids' engines are built for school reports — an adult searching a government form, a refund policy, or a medical question through a cartoon portal is using a tool against its design, and will stop within a week. So the kid-safe aisle quietly presumes the real arrangement: a children's engine bookmarked on one side, an unlocked general engine on the other, and the family's actual standard living in the gap between them. The frum need was never a sandbox next to the open web. It was one serious, complete engine, clean by construction, for everyone in the house at once — which no one on this shelf even attempts, because it requires owning an index.

The community tried, once

The desire for an owned engine is not new, and honoring the first attempt matters. In June 2009, Koogle launched — a Hebrew-language kosher search engine, built in part at the encouragement of rabbonim, omitting objectionable material, linking only to vetted news and shopping, and famously declining to operate on Shabbos. It made NPR, JTA, Ynet, ABC News. The demand it proved was real and loud. What it could not prove was sustainability: a portal skinned over someone else's index, run at side-project scale, cannot keep pace with the living web — and as of this month, koogle.co.il no longer answers at all. What remains of the owned-search dream in 2026 is earnest and narrow — eKollel's kashrus-resources directory, useful and nothing like a general engine. The lesson was never "the community doesn't want its own search." It was: an overlay can't survive; only an engine can — and nobody owned one.

What "our own engine" changes

Against that whole history, read the homepage's moat text for KolBo Search in full: "Not a filtered feed sitting on someone else's index — our own engine, built on open sources, returning clean, relevant results tuned for how this community actually searches."

Three phrases carry the architecture. Our own engine — the results page and the ranking belong to the same party that answers to this community; there is no upstream decision to contain, so the entire workaround stack — the CNAMEs, the forced settings, the screening rooms — becomes unnecessary rather than merely tighter. Clean is not a lock on the door; it is what the house is made of. Built on open sources — the index draws on the open web's raw material rather than renting another company's results feed; the exact composition is the builder's to state, and the homepage says no more, so neither do we. And tuned for how this community actually searches — the phrase every frum reader can test against their own history: the generic engines that misread "shiur times" and "cholov yisroel near me," that surround a daf yomi schedule with off-target feeds. A results page whose relevance model considers those queries normal is not a small courtesy; it is what "built for this community" means in ranking form.

One boundary, stated because this library always states them: the homepage describes the suite's AI sight protection — "state-of-the-art models screen images, video, and text in real time... protection at the level of what the eyes see" — as a property of the security layer under everything, and we quote it there, not as a Search-specific feature. The architecture speaks for itself: an owned engine under an owned enforcement layer, each doing its own job.

The front door, inside the house

The last piece is placement, and it is the piece no standalone engine — not even an owned one — could copy. On the devices this market has always known, "safe search" was a destination you forced people onto: the locked homepage, the mandated portal, the kids' engine bookmarked hopefully next to an unlocked address bar. KolBo Search ships differently: as one of the suite's "22 interoperable apps, engineered in-house, secured before they ship" — inside an operating layer where the homepage's phrase is literal: "twenty-two apps that behave like one product." The engine is the device's native front door, not a site a child must be steered toward; it sits beside the first kosher Chrome — "protection fused into the build itself" — so that the door and the map through it hold the same standard by construction. One cloud, "iOS + Android — secured on both platforms," one update pipeline, one security layer.

And the honesty beat, printed as always rather than buried: the homepage lists no standalone consumer download for KolBo Search, no public URL, and no Search-specific pricing — the suite ships via licensing to kosher device manufacturers, and the priced consumer product on the homepage is KolBo Secure, the security layer, from $14.99/month for any iPhone or Android. Where the homepage is silent — index size, language coverage, image behavior — this page is silent too; hello@kolbo.life answers what a page shouldn't guess.

Koogle proved the community wanted its own search seventeen years ago. The overlay era proved what wanting isn't enough for. The homepage's thesis line is the epitaph for both: "Anyone can remove features and call it kosher. KolBo builds what the community has been waiting for, from a blank page, to a standard the general market doesn't match." For search — the hardest surface on any device this community will ever hold — the blank page was the only place an answer could come from.

Frequently asked questions