RXed AI News

Your AI intelligence briefing
SCORE 6.3/1013/20 elements Audited 2026-07-10 · RXed table v1.0

Lovable

The fastest idea-to-working-app path a non-coder can buy — treat every output as a first draft and budget a security review before real user data touches it.

SMB toolCodingAutomation & AgentsProductivityFreemium
Vendor
Lovable · lovable.dev
Origin
SE — Stockholm
Pricing
Free (5 credits/day, max 30/mo) · Pro ~€23/mo ($25, 100 credits + 5/day, scalable to 10,000) · Business ~€46/mo ($50, adds SSO + Security Center) · Enterprise custom; PLUS usage-metered Lovable Cloud/AI billing on shipped apps ($25 Cloud + $1 AI monthly allowance included)
Users (official only)
8M users and $500M annualized revenue run rate (company-reported June 2026, up from $400M in February); 1M+ new projects/week (source, 2026-06-19)

Element scores

Reactive
Retrieval & Memory
Orchestration
Validation
Models
Primitives
Pr7
Prompts
Em
Embeddings
Cx7
Context
Tr6
Tracing
Lg7
LLM
Compositions
Fc7
Function calling
Vx
Vector store
Rg6
RAG
Gr4.5
Guardrails
Mm6
Multimodal
Deployment
Ag7
Agents
Ft
Fine-tuning
Fw8
Frameworks & harnesses
Ev3.5
Evaluations
Sm
Small models
Emerging
Ma5.5
Multi-agent
Sy
Synthetic data
Pc7.5
Protocols
In
Interpretability
Th
Thinking models
Tap or hover any element to see why it got that score.

Strengths

Nothing gets a non-technical founder from plain-English prompt to a live full-stack app faster: Build mode autonomously edits, tests in-browser and self-verifies; Lovable Cloud bundles hosting, database, storage and an AI runtime with zero external accounts; MCP support runs in both directions (server + connectors); and the generated React/Supabase code exports cleanly to GitHub. Growth is real — $500M run rate and 1M projects/week by June 2026, with Uber, HubSpot and Microsoft teams among users.

Honest dings

Security is the defining risk: CVE-2025-48757 (RLS exposure across 170+ apps), a February 2026 single-app leak of 18,697 user records, and an April 2026 BOLA flaw left open 48 days after the bug report was closed unescalated — the company's first response was denial. Credit economics punish iteration: you pay to fix the AI's own mistakes in documented debugging loops, and metered Cloud/AI billing sits on top of the subscription. The complexity wall past prototypes is real, and it is greenfield-only.

Best for: Solo founders and SMB teams validating MVPs, internal tools and demos who will harden the output (or pay someone to) before real user data arrives; not for regulated data, complex backend logic, or production apps nobody security-reviews.

Sources

Every audit lists the research it rests on — transparency and traceability are the product. Tools evolve: each audit is a snapshot of its audit date, and re-audits supersede older versions (kept below for reference).