SAASPOCALYPSEverdict #LEGORA-1E8F

scanned 2026.05.01 · 20:32

subject of investigation

legora.com

Item: legora.com
Rating: 65
Author: saaspocalypse

▸ collaborative AI legal research & drafting

verdictCONTESTED

wedge score

/100

tier · contested

wedge thesis

the door is their data moat — they're wrapping general-purpose LLMs with a legal-flavored UI, and no proprietary corpus or verified citation layer locks users in.

real walls — pick your flank·ship in 10 weeks·run for $47.00 + usage

wedge map

where the walls are.

methodology →

the door

no proprietary corpus — they're running on off-the-shelf data.

watch out

the technical wall is real — research-grade engineering, not a weekend.

capital

5.0/10

investment the incumbent had to make

technical

5.6/10

depth of the underlying engineering

network

0.0/10

users compound users

switching

4.0/10

stickiness of customer data + workflow

datadoor

0.0/10

proprietary data accumulates over time

regulatory

0.0/10

real licenses, not SOC 2 theater

take

the blunt take.

color around the thesis

“Legal AI is a crowded room of wrappers wearing suits. Legora's pitch is "collaborative" and "adapts to your ways of working" — which is product-speak for a well-designed chat interface bolted onto an LLM with some document context.”

The hard part of legal AI isn't the AI — it's verified citations, jurisdiction-aware retrieval, and hallucination guardrails that a malpractice-conscious lawyer will actually trust. None of that is proprietary to Legora, and all of it is buildable with off-the-shelf RAG tooling today.

cost

cost of competing.

their price ←→ your run-rate

what they charge●

Pro / team plan (estimated)

contact sales

/ user/mo

※ No public pricing — typical legal AI SaaS runs $50–$200/seat/mo

annual:contact sales

what running yours costs✦

01 · Vercel Pro (document uploads, edge functions)$20.00

02 · Supabase Pro (document storage + vector embeddings via pgvector)$25.00

03 · OpenAI / Anthropic API (LLM calls per review)??? — scales with usage

04 · Cloudflare R2 (PDF/doc storage)$1.00

05 · Resend (notifications, sharing)$0.00

06 · Domain$1.00

07 · OAuth (Google, Microsoft)$0.00

08 · Sentry free tier (error tracking)$0.00

TOTAL / mo$47.00 + usage

▸ break-even:immediately for any solo practitioner paying >$50/mo — your infra runs cheaper than a single seat at enterprise legal AI pricing

build

what you're up against.

est. total: 10 weeks

2 weeks document ingestion + RAG pipeline · 3 weeks review/drafting UI · 2 weeks citation verification layer · 2 weeks team/collab features · 1 week auth + billing

easy

medium

hard

nightmare

easy

Document upload + chunking pipeline

PDF/DOCX parsing with LangChain or LlamaIndex. Commodity tooling, afternoon job.

easy

Collaborative workspace UI

Shared document threads, comments, user roles. Standard SaaS CRUD with Supabase RLS.

medium

RAG retrieval tuning for legal text

Legal language is dense and clause-heavy. Hybrid BM25 + vector search needed for precision. A week of tuning.

medium

Citation grounding and hallucination reduction

Lawyers cannot cite hallucinated cases. Source-attribution with chunk-level provenance is non-trivial but solvable with careful prompt engineering + retrieval design.

hard

Jurisdiction-aware retrieval

A contract clause valid in Delaware may be void in California. Metadata tagging + jurisdiction filters across a large corpus is a real data engineering problem.

nightmare

Trust and liability positioning

Law firms won't adopt a tool that hasn't passed security review, data residency checks, and partner sign-off. Sales cycle is the moat, not the code.

stack

their position.

inferred + measured stack

detected signals· measured

cmsFramer

recommended stack · inferred

Next.js 15 + TipTap (rich doc editor)Supabase (Postgres + pgvector for embeddings)LangChain / LlamaIndex (RAG pipeline)OpenAI GPT-4o / Anthropic Claude (LLM backend)Cloudflare R2 (document storage)

rivals