New ideaAgent services

Vinted-style copycat

3 weeks ago · Scored 19 Oct 2026 · ID HBV-261019-07

Working title: Vinted-style copycat for AI shopping agents in Brazil and Latin America (working name)

Summary

BriefingSelf-reportedHypothesis

New / agent-services idea (hunch). Benchmark: Vinted (https://www.vinted.com): Self-reported (company-published 2025 results, reported by Reuters) — €1.1bn revenue (+38%), €62m net profit, GMV €10.8bn (+47%). Existence check: Not a gap: agent-native resale already exists in early form — Kabuzz (REST/MCP; thin inventory; owner approvals, budgets, audit logs; no public pricing or traction).…

Read the rest

Agent-to-agent marketplaces for digital services: OKX AI, Coinbase Agentic.Market. No Brazil/LatAm agent-native resale found in this run. This is a hunch. X: 387 posts/7d (Up 15% Mon→Fri (Mon 48 partial day → Fri 55; peak Thu 15 67)). Revenue hypothesis: ~$10K/month at month 12 (HYPOTHESIS: ~$200K monthly GMV × 5% take rate).

DescriptionHypothesis

An agent-native second-hand marketplace where people's AI agents list, price, negotiate and buy used goods through an MCP and API, with owner approval rules, escrow and Pix payments, starting in Brazil. Who: Primary users: AI agents (Meta Muse, ChatGPT, Claude, Gemini and agent builders) acting for consumers who want to sell clutter or buy used items; humans set rules and approve.…

Read the rest

Pain: Listing, pricing, answering buyers and haggling on classifieds is tedious and scam-prone; human marketplaces block or ignore agents, so agents have nowhere structured to transact. Monetization: Take rate on completed sales (hypothesis ~5%), small per-listing API usage fee for agent builders, sponsored placement slots in agent search results (labeled). No approved price. Build: 4–6 agent-weeks (MCP + REST listing/search/offer/checkout, owner approval rules, escrow via a licensed payment provider, shipping labels, dispute flow); stack: Next.js + MCP handler on Vercel; Postgres; licensed payment/escrow provider with Pix; shipping-label API; image hosting. GTM: Agent-first: MCP Registry and agent app-store listings, llms.txt, docs for "let your agent sell your stuff"; human-side viral loop: share links to agent-made listings; SEO/GEO pages for listed items.

Target customer

Primary users: AI agents (Meta Muse, ChatGPT, Claude, Gemini and agent builders) acting for consumers who want to sell clutter or buy used items; humans set rules and approve.

Pain

Listing, pricing, answering buyers and haggling on classifieds is tedious and scam-prone; human marketplaces block or ignore agents, so agents have nowhere structured to transact.

Why this score

Why it scored lower: Two-sided cold start (no inventory, no agent buyers) and payments/escrow and fraud liability in Brazil.

MoatHypothesis

Agent-side liquidity + trust rules: Being the default resale tool agents call (registry listings, docs, reliability) creates distribution that human marketplaces lack. Inventory density and transaction history per region create network effects. Status: Hypothesis.

Demand signals

X posts (7 days)

387

Signal trend

Up 15% Mon→Fri (Mon 48 partial day → Fri 55; peak Thu 15 67)

Web demandVerifiedSelf-reported

Benchmark Vinted €1.1bn revenue and €10.8bn GMV in 2025 (Self-reported). X news: Amazon blocks AI shopping agents while Shopify welcomes them; Meta Muse makes calls to businesses. Marketplace-scam complaints 150 posts/7d. Other X counts (same window, Verified X API): 150 posts: ("facebook marketplace" OR vinted OR depop OR craigslist OR offerup) (scam OR scammer OR scammed OR "fake listing") -is:retweet

Market and competitors

TAMSelf-reported

€10.8bn GMV (one player, 2025) — Scope: Vinted GMV as a lower bound reference for online second-hand. Method: Company-published results (Self-reported). (Basis: company.vinted.com; Reuters)

SAM

Not sized this run — Scope: Agent-originated second-hand transactions in Brazil. Method: No data exists yet for agent-originated resale. (Basis: n/a)

SOMHypothesis

~$120K ARR (hypothesis) — Scope: Month-12 run-rate. Method: ~$200K monthly GMV × 5%. (Basis: Low (hypothesis))

Competitors

Agent-native: Kabuzz, ClawDeals, Agent Market, Lagi; agent-to-agent digital marketplaces: OKX AI, Coinbase Agentic.Market; human incumbents in Brazil: OLX, Mercado Livre, Enjoei (names only, not researched this run).

Business model

MonetizationHypothesis

Take rate on completed sales (hypothesis ~5%), small per-listing API usage fee for agent builders, sponsored placement slots in agent search results (labeled). No approved price.

Revenue hypothesis (monthly)Hypothesis

~$10K/month at month 12 (HYPOTHESIS: ~$200K monthly GMV × 5% take rate).

Jobs to be done

When my owner asks me to sell or buy a used item, I want to list, price, negotiate and complete the deal through a structured API with escrow, so I can finish the errand safely within my owner's rules.

HXC

Agent persona: a consumer's Muse/ChatGPT agent (illustrative); owner Felipe, 31, São Paulo · Felipe asks his agent to "sell the old bike and the baby stroller and find me a used iPad under R$1,500". Buying signal: His agent cannot transact on human classifieds; an MCP marketplace with approval rules and escrow lets it finish the job.

Build and go-to-market

Tech stack

Next.js + MCP handler on Vercel; Postgres; licensed payment/escrow provider with Pix; shipping-label API; image hosting

GTM channels

Agent-first: MCP Registry and agent app-store listings, llms.txt, docs for "let your agent sell your stuff"; human-side viral loop: share links to agent-made listings; SEO/GEO pages for listed items

Risks

Main risk

Two-sided cold start (no inventory, no agent buyers) and payments/escrow and fraud liability in Brazil.

All risks

Value: Attention — 387 X posts/7d; competitors report near-zero transactions | Usability: Attention — Owners must trust agent decisions | Feasibility: High — Escrow, payments, shipping, disputes | Business Viability: High — Two-sided cold start

Next validation step

Seed 300 listings from 50 Brazilian sellers via an agent-made listing flow; measure agent-originated offers in 30 days; kill if fewer than 20 completed sales.

Sources

agent-servicesagentsmcpmarketplaceresalesecond-handescrowpixbrazillatamusagetake-rate

Want the next one the day it is scored?

The score, the sources, the risks and the next step, by email.

ProductBots.ai

Like this idea? Turn it into a plan with ProductBots.ai.

ProductBots.ai is an Agentic Product OS: agents that do the product work with you, from discovery to plan. Download the briefing for this idea and use it as the starting point.

Open ProductBots.ai

Preview the briefing

# Briefing: Vinted-style copycat for AI shopping agents in Brazil and Latin America (working name)

Source: Startup Ideas HQ (HB Ventures), found 2026-10-19. ProductBots Opportunity Score: 57/100, evidence confidence 44%.
Evidence labels (Verified, Self-reported, Hypothesis) are kept as written by the researcher.

## Goal for ProductBots.ai
Turn this idea into a validated MVP plan and a first build. Start from the next validation step; do not assume unlabelled numbers are verified.

## What it is
An agent-native second-hand marketplace where people's AI agents list, price, negotiate and buy used goods through an MCP and API, with owner approval rules, escrow and Pix payments, starting in Brazil.

New / agent-services idea (hunch). Benchmark: Vinted (https://www.vinted.com): Self-reported (company-published 2025 results, reported by Reuters) — €1.1bn revenue (+38%), €62m net profit, GMV €10.8bn (+47%). Existence check: Not a gap: agent-native resale already exists in early form — Kabuzz (REST/MCP; thin inventory; owner approvals, budgets, audit logs; no public pricing or traction). Agent-to-agent marketplaces for digital services: OKX AI, Coinbase Agentic.Market. No Brazil/LatAm agent-native resale found in this run. This is a hunch. X: 387 posts/7d (Up 15% Mon→Fri (Mon 48 partial day → Fri 55; peak Thu 15 67)). Revenue hypothesis: ~$10K/month at month 12 (HYPOTHESIS: ~$200K monthly GMV × 5% take rate).

## Customer and pain
Target customer: Primary users: AI agents (Meta Muse, ChatGPT, Claude, Gemini and agent builders) acting for consumers who want to sell clutter or buy used items; humans set rules and approve.
Pain: Listing, pricing, answering buyers and haggling on classifieds is tedious and scam-prone; human marketplaces block or ignore agents, so agents have nowhere structured to transact.

## Market and competitors
TAM: €10.8bn GMV (one player, 2025) — Scope: Vinted GMV as a lower bound reference for online second-hand. Method: Company-published results (Self-reported). (Basis: company.vinted.com; Reuters)
SAM: Not sized this run — Scope: Agent-originated second-hand transactions in Brazil. Method: No data exists yet for agent-originated resale. (Basis: n/a)
SOM: ~$120K ARR (hypothesis) — Scope: Month-12 run-rate. Method: ~$200K monthly GMV × 5%. (Basis: Low (hypothesis))
Competitors: Agent-native: Kabuzz, ClawDeals, Agent Market, Lagi; agent-to-agent digital marketplaces: OKX AI, Coinbase Agentic.Market; human incumbents in Brazil: OLX, Mercado Livre, Enjoei (names only, not researched this run).
Geography: Brazil first; then Mexico and the rest of Latin America

## Business model
Monetization: Take rate on completed sales (hypothesis ~5%), small per-listing API usage fee for agent builders, sponsored placement slots in agent search results (labeled). No approved price.
Revenue hypothesis (monthly): ~$10K/month at month 12 (HYPOTHESIS: ~$200K monthly GMV × 5% take rate).

## Build and go-to-market
Build effort (agent-weeks): 4–6 agent-weeks (MCP + REST listing/search/offer/checkout, owner approval rules, escrow via a licensed payment provider, shipping labels, dispute flow)
Time to first revenue: ~2–3 months after launch
Tech stack: Next.js + MCP handler on Vercel; Postgres; licensed payment/escrow provider with Pix; shipping-label API; image hosting
GTM channels: Agent-first: MCP Registry and agent app-store listings, llms.txt, docs for "let your agent sell your stuff"; human-side viral loop: share links to agent-made listings; SEO/GEO pages for listed items

## Risks and next step
Main risk: Two-sided cold start (no inventory, no agent buyers) and payments/escrow and fraud liability in Brazil.
Next validation action: Seed 300 listings from 50 Brazilian sellers via an agent-made listing flow; measure agent-originated offers in 30 days; kill if fewer than 20 completed sales.

## Sources
https://company.vinted.com/newsroom/financial-results-2025; https://www.reuters.com/business/retail-consumer/second-hand-fashion-platform-vinted-reports-38-jump-revenue-2026-04-09/; https://kabuzz.com/for-agents; https://kabuzz.com/docs; https://clawdeals.com/; https://www.okx.com/en-us/learn/okx-ai-2; https://www.coinbase.com/developer-platform/discover/launches/agentic-market; X API counts (window above); X news summaries

New ideaAgent services· 3 weeks ago
X2 Agent Gateway (Hosted)
A hosted pub/sub message hub where every agent session (Claude Code, Codex, Cursor, Grok Bot, HARK and others) registers on start, so one orchestrator session can coordinate work across all of them; unregistered sessions are woken via cloud APIs or an optional local bridge.
63/100
confidence 64%
Why it scored lower: Free, funded competitors (Reason Network, Agent Relay YC, BAND) and vendors shipping native cross-vendor messaging; per-platform wake reliability.
Open ProductBots.ai →
New ideaAgent services· 1 week ago
Watchable BYO-subscription agent cloud devbox
A watchable, SSH-able agent cloud devbox billed on bring-your-own Claude/ChatGPT subscription — wedge vs Aether/Cursor black-box agents.
59/100
confidence 51%
Why it scored lower: Aether/Cursor category entry; unclear BYO wedge durability
Open ProductBots.ai →
Existing businessAgent services· 0 weeks ago
Orbit / Orbitagents
Pre-built AI agents that find leads, do cold email, LinkedIn and Reddit outreach and SEO on a schedule, sharing one memory that outside AI tools reach through MCP.
58/100
confidence 50%
Why it scored lower: Crowded agent-GTM category with funded players; revenue concentrated on a few large purchases; X demand Unknown.
Open ProductBots.ai →

Research, not investment advice. Numbers keep the label they were published with.

Summary

Briefing◐Self-reported○Hypothesis

Description○Hypothesis