Legacy platforms
Relativity · DISCO
- Built pre-LLM; AI bolted on as an add-on.
- Enterprise pricing and enterprise onboarding.
- Processing fees, user fees, hosting fees, still.
- Weeks to implement; months to train a team.
AI workspace for legal matters
Open a matter. Ask anything. Cite everything.
A research chat and a review table, grounded in the files of one matter. Every answer shows the exact excerpt it came from.
What we promise about your data
-
No training
Your matters never leave your workspace.
Documents and prompts are never used to train our models or anyone else's. Not now. Not later. Written into the DPA.
-
Matter-scoped
Isolation is the default, not a setting.
Every matter is walled off. No cross-client mixups, no lateral reads, no shared embeddings. Role-based access per matter.
-
Your data, your rules
Export everything. Delete anytime.
CSV for reviews, original files for documents, a full audit log for the whole workspace. Workspace deletion is immediate and irreversible.
I. The default mode
Ask with context. Get answers with receipts.
Every matter has its own research chat, grounded in the files you pick and filtered by the tags you set. Pull the web in when you need current law. Notes sit one tab away. Every answer lands with numbered citations; click one and the exact source excerpt opens inline.
II. When you need scale
One question. Every document. Answered in parallel.
Review is a spreadsheet where every column is a question (text, yes/no, or date) and every row is a document in the matter. The AI fills the grid. You audit the edge cases and export to CSV.
III. The flow
Open. Ask. Review.
-
Step 01
Open a matter.
Upload files, tag them, add notes. Invite collaborators with per-matter permissions. Every matter is its own walled-off workspace.
-
Step 02
Ask with context.
Chat with the files you choose. Filter by tag or folder. Add web search when you need current law. Every answer shows the source excerpt it came from.
-
Step 03
Review at scale.
When you need the same answer on every document, build a review table. Columns are text, yes/no, or date. The AI fills the grid; you audit and export to CSV.
IV. The honest comparison
Built for the work, not for the enterprise sale.
Outsourced review
Contract review teams
- $0.50–$2.00 per document, per pass.
- Turnaround measured in weeks.
- Consistency varies reviewer to reviewer.
- Client documents leave the firm.
LawGPT
AI-native. Self-serve.
- For lawyers, by lawyers. Designed by people who've done the first-pass themselves.
- Built on LLMs from day one, not bolted to them.
- Transparent credit pricing. $170/month to start. No per-document fees.
- Open a matter in minutes. First review in seconds.
- Self-serve: sign in with email or Google and go.
V. The ones you'll ask
Answers before the email.
Legal buyers have the same six questions in every demo. Here they are. Billing-specific ones live on the pricing page.
Is this privileged?
We treat every document as confidential by default. Humans on our team don't access your content except during a support request you initiate, or if we're investigating a concrete security incident involving your matter. Attorney-client privilege isn't a product feature; we operate on the assumption that anything you upload may be privileged.
Do you train AI models on our data?
No. Documents, prompts, and outputs are never used to train, fine-tune, or evaluate models, whether ours or a third party's. Inference runs against providers under agreements that prohibit training on API inputs, and we pass the same commitment through to you in the DPA.
Do I need to enter a card to get started?
Yes. Sign up with email or Google, create your workspace, and pick a plan. Stripe Checkout collects the card before your first billable action runs. Your monthly subscription credits cover chat, review, summaries, and storage; everything in the product bills against that one balance.
Can my whole firm use one workspace?
Yes. Workspace roles are owner, admin, member, and external. Matter-level roles are owner, editor, viewer, and external. An admin can invite a collaborator to a specific matter without giving them access to the rest of the workspace. Seats are included in each plan's allowance.
What happens to my documents if I cancel?
You keep access through the end of the paid period and can export reviews to CSV at any time. After cancellation your workspace and its content stay in place — reactivate the subscription whenever you're ready and pick up where you left off. To permanently delete a workspace, an owner triggers deletion from workspace settings; that action is immediate and irreversible.
Do we still need Relativity or DISCO?
If you're running billion-document productions with chain of custody requirements for regulators, yes, keep them. If you're a small or mid-size firm doing first-pass review on matters of a few thousand to a few hundred thousand documents, LawGPT replaces the first pass entirely, and you'd use eDiscovery platforms only for production cuts.
VI. Pricing
Two plans. One unit of usage.
A monthly credit allowance covers chat, review, summaries, and storage. One balance, one rate card.
Standard
Solo & small firms
$170 / month
- 1,500 credits / month
- ≈ 1,500 chat messages or review cells
- ≈ 500 AI summaries
- ≈ 7 GB stored for a month
Premium
Mid-size firms reviewing at volume
$700 / month
- 7,500 credits / month
- ≈ 7,500 chat messages or review cells
- ≈ 2,500 AI summaries
- ≈ 37 GB stored for a month
VII. Security & data handling
What happens to your data. In plain English.
- We don't train on your data. Ever.
- Your documents, prompts, and outputs are never used to train our models or any third-party's. Written into every DPA.
- Matter-scoped by default.
- Workspace roles (owner, admin, member, external) plus per-matter permissions. No embeddings shared across matters. Audit log on every meaningful action.
- Your data leaves when you do.
- Export reviews to CSV, documents in their original formats, and the audit log at any time. Workspace deletion happens the moment an owner confirms it — there is no grace period.
- Plain English over checkbox security.
- SOC 2 attestation is on the roadmap, not done yet. We'll share the current control posture and prior pen-test results under NDA.
- US data residency.
- Primary storage, backups, and indices all stay in US regions on Azure. TLS everywhere, AES-256 at rest, managed keys.
- DPA, BAA, and your paper.
- Standard DPA and BAA available on request. Custom redlines go through a real human on our side, not a legal-ops queue.
Model training
Access & isolation
Retention & export
Compliance posture
Infrastructure
Legal
Full details on the security page.
VIII. Your next matter is waiting.
Pick a plan. Open a matter.
Sign up, choose Standard or Premium, and bring your documents, or start with a sample matter we've prepared.