Lemrock
Also known as: Lemrock AI, Lemrock AI SAS
Paris deeptech building vendor neutral middleware for agentic commerce: a single integration ingests a brand's catalog, structures it for large language models, and makes products discoverable and transactable inside AI assistants like ChatGPT, Claude, Gemini, and Perplexity while brands keep control of pricing and margins.
Lemrock is a Paris deeptech startup building middleware for agentic commerce, founded in 2025 and incorporated in December 2025 by Roxane Laigle, a former Fnac Darty Strategy Director, alongside Sasha Collin and Clement Nguyen, who previously co founded the Y Combinator backed startup Mindely. It raised a six million euro, roughly seven million dollar, seed round in March 2026 led by Galion.exe, with Criteo founder Jean-Baptiste Rudelle joining the board.
The thesis is a structural shift in how people shop: as consumers move from Google searches to asking ChatGPT, Claude, Gemini, or Perplexity which product to buy, some retailers already report website traffic declines of up to thirty percent, and the infrastructure for brands to be visible and transactable inside those conversations does not exist at scale. Lemrock is that missing layer. Through a single integration it ingests a brand's product catalog, structures it so large language models can retrieve, interpret, and surface it inside conversations, exposes real time pricing and availability, executes transactions, and tracks performance, closing the loop between recommendation and purchase. Its stance is deliberately vendor neutral and brand safe: one integration reaches multiple AI platforms while brands keep control over pricing, margins, and product representation. Because conversational agents gather context, history, and user preferences during a conversation, Lemrock uses that information to sharpen recommendations in real time with a granularity traditional search cannot match, powered by proprietary models trained on millions of data points. In under a year it has onboarded more than sixty brands including Maisons du Monde, Cdiscount, Darty, Leroy Merlin, Lidl, and Rakuten, processing more than one hundred million interactions monthly.
Lemrock fits European and US retailers and brands that want to appear and sell inside AI assistants without building bespoke per platform integrations. As a young, narrow middleware layer it is a weaker fit for buyers needing published pricing, documented security certifications, or a broad agent development platform, and it faces the long term risk of AI platforms building native commerce features.
Vendor details
Canonical URL
https://lemrock.ai
Category
Agent infrastructure
Subcategory
Agentic commerce catalog and transaction middleware
Funding status
Independent, founded in 2025 near Paris and legally incorporated as Lemrock AI SAS in December 2025, by Roxane Laigle, CEO and former Strategy Director at Fnac Darty, Sasha Collin, CPO, and Clement Nguyen, CTO, with Collin and Nguyen previously co founders of the Y Combinator backed startup Mindely. Raised a six million euro, about seven million dollar, seed round announced March 11, 2026 led by Galion.exe, with Criteo founder Jean-Baptiste Rudelle joining the board and angels including Michael Benabou of Veepee, Gary Anssens of Alltricks and Decathlon, and Antoine Lizee of Alan, plus consulting firm Sia. Selected for the Agoranov deeptech incubator and named to the Forbes 30 Under 30 Europe list.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
A single integration ingests a retailer's product catalog including pricing, availability, and attributes and structures it for large language models, then distributes it across multiple AI platforms including ChatGPT, Claude, Gemini, and Perplexity. Handles real time price and availability exposure, transaction execution, and performance tracking through one orchestration layer, and the company reports processing more than one hundred million AI commerce interactions per month across ten million plus live products.
In practice
A shopper asks ChatGPT which sofa to buy. Lemrock has structured a retailer's catalog so the right products surface with correct real time pricing and availability, and the purchase can complete inside the conversation.
A brand sees website traffic fall as customers shift to AI assistants. Rather than building a separate integration for each platform, it connects once to Lemrock and becomes visible and transactable across ChatGPT, Claude, Gemini, and Perplexity.
A retailer worries about losing control of pricing and margin inside AI channels. Lemrock keeps the brand in control of pricing, margins, and product representation while exposing the catalog to agents.
Sources & related URLs
Related / legacy domains
Research sources
Capability coverage
5.5 / 14 capabilities · 39%
| Integrations & Tool CallingA single integration ingests a retailer's product catalog with pricing, availability, and attributes and distributes it across ChatGPT, Claude, Gemini, and Perplexity, exposing real time data and enabling transactions, EU-Startups and Sia consultancy coverage retrieved 2026-07-08 | Full |
|---|---|
| Workflow OrchestrationA complete middleware stack connects product catalogs, AI agents, and transactions through a single orchestration layer, closing the loop from recommendation to purchase across multiple ecosystems, Sia consultancy and EU-Startups coverage retrieved 2026-07-08 | Full |
| Knowledge Grounding & RAGIngests and structures product catalog data so large language models can retrieve, interpret, and surface it inside conversations, backed by proprietary models trained on millions of data points, Sia consultancy and TAMradar coverage retrieved 2026-07-08 | Full |
| Human Oversight & GuardrailsThe platform is brand safe by design, letting brands keep control over pricing, margins, and product representation, though a documented human approval workflow is not described, TheNextWeb and Sia consultancy coverage retrieved 2026-07-08 | Partial |
| Security, Identity & GovernanceThe platform handles transactions and catalog data and references secure transaction processing, but no security certifications or governance controls are documented on retrieved pages, inforcapital and EU-Startups coverage retrieved 2026-07-08 | Unable to verify |
| Observability & AuditabilityThe platform tracks performance and provides real time insights on brand presence inside AI channels, though agent run tracing and audit trails are not documented, Sia consultancy and TAMradar coverage retrieved 2026-07-08 | Partial |
| Memory & State PersistenceConversational agents gather context, history, and user preferences during a conversation, and Lemrock uses that information to improve recommendations in real time, though a persistent agent memory architecture is not documented, EU-Startups coverage retrieved 2026-07-08 | Partial |
| Deployment & Data ResidencyDelivered as a cloud middleware service with no self hosted or data residency options documented, EU-Startups coverage retrieved 2026-07-08 | Unable to verify |
| Prebuilt Agents, Templates & PacksLemrock is catalog and transaction middleware that makes brands available to AI agents rather than shipping prebuilt agents or templates, EU-Startups and TheNextWeb coverage retrieved 2026-07-08 | Unable to verify |
| Triggers & Channel CoverageDistributes across multiple AI assistant channels including ChatGPT, Claude, Gemini, and Perplexity so products surface when users query them, though a traditional inbound event trigger framework is not documented, EU-Startups and Forbes profile retrieved 2026-07-08 | Partial |
| Model Flexibility & RoutingLemrock runs proprietary models internally and distributes across third party AI platforms as channels, but the customer does not choose or route the underlying models, TAMradar coverage retrieved 2026-07-08 | Unable to verify |
| APIs, SDKs & MCP ExtensibilityBrands connect through a single integration point to structure catalogs and reach multiple AI platforms, though a public developer SDK or MCP surface is not documented, EU-Startups and TheNextWeb coverage retrieved 2026-07-08 | Partial |
| Testing, Debugging & OptimizationNo testing, evaluation, or optimization tooling is documented on retrieved pages, EU-Startups coverage retrieved 2026-07-08 | Unable to verify |
| Browser & Computer UseNo browser or computer use capability is described on retrieved pages, EU-Startups coverage retrieved 2026-07-08 | Unable to verify |
Pricing
No public pricing; contracts are quoted through sales
not published; B2B contract likely scaled by catalog size, interaction volume, or transactions
What is public
Nothing numeric. The B2B model, customer roster, and scale metrics are public, but no rates are published.
Billing mechanics
B2B sales led motion aimed at brands and retailers that cannot build the integration stack in house. As an early stage company incorporated in December 2025, commercial packaging is likely still forming.
Cost watchouts
As a transaction and distribution layer, any per interaction or per transaction component would scale with the hundred million plus monthly interactions the platform already processes for large catalogs.
Variable cost rationale
No public rates exist. As a distribution and transaction middleware processing more than one hundred million interactions monthly across large catalogs, cost could carry a per interaction or per transaction component that scales with usage, but the company publishes no mechanics, so treat exposure as unquantified and quoted per deal.
Additional watchouts
This is a very young company with pricing likely to evolve; confirm terms directly and clarify whether any transaction or volume based component applies given the scale of interactions.
Sales call required
Yes — required for paid access
Free / trial
No free tier or trial is documented on retrieved pages
Key ambiguities
Whether pricing is a platform fee, per interaction, per transaction, or a share of sales driven through AI channels.
Missing data
No published rates, tiers, minimums, or contract terms were retrievable from company or third party pages.
Related vendors
- Acrab — Singapore compute infrastructure company building a full stack…
- AgentOps — Agent observability and reliability platform with broad model and…
- Agno — High-performance agent runtime and framework (formerly Phidata) with…
- AIsa — Unified resource and payment gateway for AI agents that lets them…
- Apify — Cloud platform for web scraping and automation with 45,000+ prebuilt…
- Arcade — Authenticated tool calling platform and MCP runtime that handles…