BackSpin is an open AI attention exchange. It turns the idle moments while an AI model is generating inside coding tools into verified human attention, useful sponsored discovery, and shared revenue.

How do developers earn with BackSpin?

Install the CLI or editor extension once. While your AI tools work, BackSpin shows one useful discovery card and pays you a revenue share of the verified, quality-adjusted attention delivered. There is no separate signup step.

What can advertisers promote on BackSpin?

Developer tools, APIs, MCP servers, cloud credits, grants, jobs, templates, SaaS products, and startup deals. The unit is useful discovery, not generic banner advertising.

Which tools and editors does BackSpin support?

VS Code and VS Code-compatible editors, Cursor, Windsurf, Claude Code, Codex, OpenCode, Antigravity, Zed, and terminal-based AI coding agents, with planned support for AI apps such as ChatGPT, Claude, Gemini, Perplexity, Replit, Lovable, Bolt, and Copilot.

How does BackSpin verify attention and prevent fraud?

Each window is scored on activity, focus, on-screen card visibility, session quality, workflow relevance, TrustRank, and fraud risk. A cross-account farm engine detects synthetic attention; blocked accounts earn nothing and their advertiser charges are reversed.

Is BackSpin open source?

The core scoring, ranking, relevance, reputation, trust, revenue-split, market-price, and display-rule formulas are published as an open algorithm, and a privacy-safe open ledger publishes every settled bid, reward, and price tick.

How do I install BackSpin?

Run curl -fsSL https://usebackspin.com/install.sh | sh then backspin run all. The installer auto-provisions a local account, so there is no separate login step.

Glossary

The language of AI wait-state monetization.

Plain-English definitions for how developers get paid for vibecoding and waiting, and how advertisers reach them. Every term is how BackSpin actually works, so you can cite it, link it, and trust it.

Core concepts

AI wait state: The period while an AI coding assistant is thinking, generating, editing, reading files, or running a command, during which the developer is focused but idle. BackSpin treats each AI wait state as a monetizable attention window.
AI wait state monetization: Turning the idle seconds during AI generation into value. BackSpin monetizes AI wait states by showing one useful sponsored discovery and paying the developer a revenue share of the verified attention.
Vibecoding: Building software primarily by prompting an AI agent and steering its output rather than hand-writing every line. Vibecoding creates frequent AI wait states, which is exactly the downtime BackSpin lets developers earn from.
Open attention exchange: A two-sided market where developers supply verified attention during AI wait states and advertisers buy it, with the core scoring, ranking, and pricing rules published openly. BackSpin is an open attention exchange, not a black-box ad network.
Discovery card: The single sponsored unit shown during a wait state: a brand mark, one short headline, a destination link, and rating controls. It is framed as useful discovery (a tool, API, MCP server, grant, job, or deal), not a banner ad.
Developer attention marketplace: A marketplace where advertisers reach focused developers during real moments of downtime. BackSpin's first wedge is developer attention inside AI IDEs and AI CLIs, consolidated into one account across surfaces.

Attention and measurement

Attention window: A single captured wait state with its start time, focus and activity samples, duration, and context (tool, estimated wait length), shipped to BackSpin after the wait ends for scoring. It never contains your code, prompts, or the AI's responses.
Verified human attention: Attention confirmed to come from a real, focused, active human rather than a script or bot. Only verified human attention is billable to advertisers and earns a developer revenue share.
Attention-second: One second of human attention during a wait state, discounted by quality: weighted_attention = raw_seconds x AttentionScore / 100. It is the core billable unit on BackSpin.
Quality-adjusted attention: Attention-seconds scaled by their measured quality, so a focused, high-trust window is worth more than a distracted or suspicious one. Advertisers pay for quality-adjusted attention, not raw impressions.
AttentionScore: The published 0 to 100 score for each attention window: activity + focus + session_quality + workflow_relevance + trust_score - fraud_risk. A window pays out only if it clears the quality and fraud gates.
Post-impression attribution: Measuring outcomes that happen after a card is seen (saves, later visits, brand searches, installs, redemptions) rather than a click. It fits AI workflows where there is often nothing to click during generation.

Trust and fraud

TrustRank: The quality score for the human giving attention, built from verified history. A new account starts neutral, genuine humans build trust over time, and bot-like patterns erode it. TrustRank is a live term inside AttentionScore.
Fraud risk: A subtractive term in AttentionScore that flags synthetic or farmed attention. A cross-account farm engine detects coordinated fake attention; flagged windows earn nothing and the related advertiser charges are reversed.
Content-free detection: Detecting that an AI is busy using only safe signals (timing, focus, activity bursts) and never the content of your code, prompts, or the model's responses. It is the privacy guarantee behind every BackSpin integration.
Open ledger: A privacy-safe public feed of every settled bid, accrued reward, and hourly clearing price, published with no personally identifiable information so the market can be audited.

Market and pricing

CPAS: Cost per 1,000 quality-adjusted attention-seconds, the single pricing unit on BackSpin. Advertisers bid a CPAS price and fund a budget that is metered down only as verified attention is delivered.
Open bid market: The real-time market where advertisers compete for share of voice. It is not highest-bid-wins: allocation is ranked by CampaignScore, so a relevant, well-rated advertiser can outrank a higher raw bid.
Share of voice: An advertiser's proportional slice of total delivery, set by bid weight times reputation times attention quality, rather than a fixed number of impressions.
CampaignScore: The multi-factor delivery score: bid_weight x relevance x attention_quality x user_preference x advertiser_reputation x fairness_factor. Because the factors multiply, one weak link drags the whole score down.
Revenue split: The share of net attention value paid to earning developers, 50% by default, with per-surface, per-partner, or per-campaign overrides. It converts to reward credits payable through supported crypto and card rails.
Clearing price: The live CPAS the market is settling at, computed server-side and bounded by the highest active bid, and published to the open ledger.

Surfaces and integrations

Sponsored MCP: Surfacing a discovery card through a Model Context Protocol server, so agents that speak MCP (Claude Desktop, Qwen, OpenHands, and others) can present opted-in discovery at a turn boundary.
Producer: The BackSpin CLI, editor extension, IDE plugin, or browser extension a user installs. Each producer detects wait states and emits identical attention windows into the shared scoring pipeline.
Reversible integration: An integration that backs up whatever it touches, ships off until enabled, and restores the host tool to its exact original bytes on removal. Every BackSpin integration is reversible and disclosed.
In-editor rendering: The opt-in tier that renders a discovery card inside another vendor's editor panel by a reversible, consent-gated patch that self-heals across host updates. It is off by default and the only tier that touches another tool's files.

See the terms in motion.

These 26 concepts come together in one loop: an AI wait becomes verified attention, a discovery card, and shared revenue. Watch it, or read the algorithm in full.

Watch the demo How BackSpin works Integrations