PeerPush Signals

The major AI assistants read PeerPush product data around the clock, and a community of real builders votes the best tools up. Here is the scale of both.

trailing 30 days, updated daily
Product reads by AI assistants
all time
Upvotes from real builders
trailing 30 days, updated daily
AI systems reading PeerPush

Where software gets built

Every country builders ship from, sized by the work that clears the quality bar and coloured by how much of that country's output clears it.

Read by the AI systems people actually use

Eight weeks of product reads, stacked by engine. Press play to watch the lead change hands.

How we measure this

AI visibility

PeerPush exists to be readable by machines. This is how far that reaches.

Read by every major AI system

Three separate ways AI finds a product, each measured on its own signal over the last 8 weeks.

In the index

Crawled and stored, so a product can be recalled in any later answer.

share of listed products read

Pulled for a live question

Fetched in the moment, while a person is asking their assistant.

share of listed products fetched live

Sent a person

Someone followed a link out of a chat answer onto the product page.

share of AI-chat visitors

The shape of the market

What people look for and cannot find, and which capabilities keep shipping together.

What people cannot find

How much people ask for each theme, against how well the catalog answers it, last 30 days.

What ships together

Use cases that keep turning up in the same product, drawn as the links between them.

Trust and momentum

Which categories are moving on trust, conviction, shipping and discussion.

Market pulse

Categories ranked by peer trust, community conviction, public shipping or discussion. Switch the metric to change what the ranking means.

Frequently asked questions

PeerPush Signals is a free market-intelligence layer that shows how AI systems read the software market, where software gets built, what people look for and cannot find, and where trust and momentum are moving, measured from both human and AI activity on PeerPush.
It tracks which AI systems read PeerPush and how far into the catalog each one reaches, where software is built and how much of each market clears the quality bar, the search themes the catalog answers badly, which use cases keep shipping together, and how categories rank on peer trust, conviction, public shipping and discussion.
Signals are derived from real traffic and engagement on PeerPush: AI crawlers, AI agents browsing on behalf of users, MCP tool calls, and human visitors. Only privacy-safe, aggregated shape is published.
Aggregations refresh on rolling schedules: short windows update every few minutes and longer windows on slower schedules. Each page shows a "Data fresh as of" timestamp.
AI traffic combines AI crawlers, AI agents, and MCP clients, classified by user-agent and entry point. The methodology page explains exactly what is counted and what is excluded.
Yes. Everything on the public Signals surfaces is free to read, and every published figure passes a k-anonymity floor so no small audience can be identified.

How we count AI traffic

What “AI Surface Area” combines, how we treat redaction, and why we don't show absolute counts until volume crosses a floor.