Opportunities › Programmable internet search and indexing › Proposed app
Rowlight
Query a web-scale index with SQL and Datalog. Bid for speed, free when idle.
- Score
- 42
- confidence 68
- Evidence
- 11
- items on the page
- Rivals
- 3
- named in sources
- First version
- 60 d
- estimate
- Type
- Desktop app
- PWA: runs in the browser, installs on Mac and Windows
- Market since
- 2021
- Emerging market · AI estimate
ScreensConcept
Desktop app: Writing and tuning SQL or Datalog queries, previewing costs, and reviewing results is sustained technical work requiring a keyboard and full focus. Sample screens on a computer, in the proposed brand. Their numbers are sample data.
New query
Free at current load$0.00
About this appAI proposal
Rowlight is a search index you query directly. Instead of sending text to a closed search API, you or your AI agent write SQL or Datalog against the index and decide how results are filtered, joined and ranked.
It is for AI builders, agent developers and information product makers who want to see and change their search logic. When capacity is free a query costs nothing. At busy times you place a small bid for priority.
Different from fixed-price search APIs: the ranking logic is yours, query recipes can be shared and forked, and a hard query is not priced the same as an easy one.
- Tagline
- Write the search. Pay only what the load costs.
- Voice
- PlainPreciseOpen
- Why this name
- Rowlight shines a light on rows: the index is a table you can read, not a black box. Short, plain and easy to say.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Direct SQL and Datalog access
Write queries against the index yourself. Filters, joins and ranking are in your hands, not hidden behind an endpoint.
No member yet
Congestion-based micro-auctions
Set a maximum bid per query. When load is low the price is zero; when busy, higher bids run first.
No member yet
Agent-ready query drafting
Describe what you need and an LLM drafts the SQL. You review, edit and run it, or let your agent call the API.
No member yet
Cost preview before you run
See estimated scan size and the current clearing price first, so hard queries never surprise you.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#1E5B4FDeep Pine · primary#F6F3ECPaper · surface#1F2A24Ink · text#E07A2FEmber · accent#3A6EA5Slate Blue · data
Typography
Aa
Rowlight
ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
Write the search. Pay only what the load costs.
PricingAI proposal
Exa and Tavily list paid plans with prices on request plus a free plan, so the material gives no competitor prices. These tiers are a starting point to test, not derived from them.
Idle
Free
- Queries run free when capacity is available
- Non-commercial use
- Public recipes
Builder
$20 a month
- Bid credits included
- Priority bidding at busy times
- API keys with spend caps
- Private recipes
Scale
Contact us
- Reserved capacity
- Higher rate limits
- Support and invoicing
Would you pay Builder · $20 a month?
Your answer tells PWAS which prices people accept. It changes no score.
No answers yet.
What it could be worthEstimate
What a software business making this much sells for today, from real deals. Pick a monthly revenue.
If it makes
Worth about
$194,000
Sale range $132,000 to $263,000
$120,000 a year in revenue
From 87 software businesses listed and sold on Empire Flippers in the last year, revenue vetted by the broker: they sell for about 19× their monthly revenue, most between 13× and 26×.
Measured on profit, the middle deal is 32× monthly net profit: about 2.7 years of profit.
An estimate, not an offer or financial advice. A real price depends on profit, growth, churn, how much of the work is the owner's, and the buyer.
How it would growAI proposal
How the product would reach new customers through its own use, and how the first ones would be found. A plan to test, not a result.
- Built-in loop
- Shared outputs. Users publish query recipes and share result links. Each public recipe page shows the product name and a Fork button, so readers can run it themselves.
- Do they become customers?
- Readers of a recipe are often other builders who could sign up, but a link viewer may only look. Test how many recipe visitors run a query and then create an account.
- First customers
- Solo and small-team developers building AI agents or research tools that already call a search API and are comfortable writing SQL.
- Where to reach them
- Hacker News: Share a demo with live queries and the auction price; the idea already drew 61 points and 26 comments there.
- Agent developer communities: Post recipes for common agent tasks in forums and Discords where builders compare search APIs.
- First milestone
- Fifty weekly active builders running queries, and ten paying for priority bids within eight weeks of launch.
Why people already pay for this
- VerifiedHacker News post received 61 points and 26 comments
- Public claimExisting search APIs have opaque algorithms with no user control and fail to fairly distribute compute resources across queries of varying difficulty.
- Public claimUsers cannot improve or customize search strategies with existing products, limiting community-driven improvements.
- Public claimExisting search companies have abandoned text-to-SQL approaches, leaving an opportunity for programmable SQL-based search.
Tech stack
Every proposal builds on the same PWA base, installable from the browser on any phone or computer. Dashed boxes are what this app adds.
- AppNext.js and TypeScript
- PWAManifest and service worker
- UITailwind CSS
- Data and authSupabase Postgres
- PaymentsStripe
- EmailResend
- HostingVercel
- This app · Index storeClickHouse cluster with NVMe storageHolds the web-scale index and runs user SQL queries fast.
- This app · Query sandboxQuery limits, quotas and a Datalog engineStops costly or abusive queries and supports Datalog besides SQL.
RisksAI inference
- TechnicalServing untrusted SQL over a very large index is hard: one heavy query can starve others, so limits and fair auction scheduling must work from day one.
- DataBuilding and refreshing an internet-scale index costs a lot in crawling and storage, and index quality decides whether builders switch from existing APIs.
- CommercialBuilders may prefer a simple search call to writing queries, and variable auction prices may be harder to budget than fixed plans.
- Intellectual propertyCrawled content may carry copyright and terms-of-use limits on what can be stored and returned to commercial users.
- AILLM-written SQL can be wrong or expensive; the cost preview and caps must catch this before agents run it.
Everything marked AI proposal or AI inference was written by a model from the evidence on the opportunity page. The proposed name must be checked against trademark registers before it is used.