01Launched: December 2025

UFound

AI matching for startup teams and co-founders

How it is built

37 technologies in 7 layers

On top is what the user sees. Underneath is everything that keeps the product running. Every technology here is in use right now.

UFound: AI matching for startup teams and co-founders

InterfaceTypeScriptReactNext.jsCSS3ReduxPlaywrightVitest
Server sideNode.jsFastifyRabbitMQBullMQ
DataPostgreSQLRedisElasticsearchMinIO
IntegrationsTelegram Bot APIWebhooks, file exchange
AIClaudeGeminiDeepSeekLangGraphQdrant
InfrastructureDockerKubernetesHelmIstiocert-managerVaultArgoCDGitLab CINginxLinux
MonitoringPrometheusGrafanaLokiOpenTelemetryJaeger

What it solves

Before: Putting a team together drags on for six months or more, and the project sits still the whole time.
Now: The algorithm returns recommendations within a day of the profile being filled in: the search starts from a shortlist, not from an empty screen.
Before: The CVs line up, but working together does not: different expectations and a different pace.
Now: Compatibility is scored across 40+ parameters: skills, experience, working style, equity expectations and motivation. The result is shown broken down by each of them separately.
Before: A specialist cannot tell a live project from an intention, and risks walking away with nothing.
Now: The project profile spells out the terms and the equity before the conversation starts, and verification and portfolios separate confirmed participants from anonymous ones.
Before: Equity is agreed verbally and falls apart at the first disagreement.
Now: NDA templates and terms recorded in the open move the agreement out of a chat thread and into a document.

12 capabilities

All of this had to be designed, built and taken into production.

Matching

Two-sided algorithmCompatibility is scored for both sides at once: the founder and the specialist each see how well the person suggested to them fits.
40+ parametersSkills, experience, working style, equity expectations and motivation, rather than one keyword overlap in a CV.
Recommendations within 24 hoursThe first shortlist arrives within a day of the profile being completed.
Compatibility breakdownThe result opens up by area: skills separately, expectations separately, working style separately, so you see where you converge and where you do not.

Trust and terms

Terms and equity in the openWhat is on offer and on what terms is visible in the project profile, before the first message.
Verified participantsA confirmed status separates vetted participants from anonymous profiles.
NDA templatesThe legal frame is ready in advance, so the conversation does not stall on "somebody should draw up a document".
Portfolios and proofWork is attached to the profile and confirmed, so claimed experience stops being just words.

Working on the project

Built-in messengerConversations, history and attachments inside the platform: agreements do not get lost across three channels.
Progress trackingThe state of the project and each person's contribution are visible on a shared board.
Detailed project profileStage, tasks and roles are set out so a specialist understands the scope before the conversation.
Privacy controlsVisibility of the profile and contacts is configurable, so searching does not mean going public.

Need the same?We build this for you too:

Product plans

These are the product's own prices, not the cost of development.

Basic0 ₽Unlimited projects, search, filters and AI recommendations.
Pro990 ₽ / monthPriority in search, confirmed status and an extended project description.
Business2 990 ₽ / monthPermanent project visibility and priority placement.

Message us on Telegram

What it solves

Before: Putting a team together drags on for six months or more, and the project sits still the whole time.

Now: The algorithm returns recommendations within a day of the profile being filled in: the search starts from a shortlist, not from an empty screen.

Before: The CVs line up, but working together does not: different expectations and a different pace.

Now: Compatibility is scored across 40+ parameters: skills, experience, working style, equity expectations and motivation. The result is shown broken down by each of them separately.

Before: A specialist cannot tell a live project from an intention, and risks walking away with nothing.

Now: The project profile spells out the terms and the equity before the conversation starts, and verification and portfolios separate confirmed participants from anonymous ones.

Before: Equity is agreed verbally and falls apart at the first disagreement.

Now: NDA templates and terms recorded in the open move the agreement out of a chat thread and into a document.