Skip to content

Contract check + negotiation prep

Know what you're signingbefore you negotiate it.

Paste a term sheet, job offer, freelance agreement or SaaS renewal. Get every clause that costs you money — with the market norm and the wording to ask for — in seconds. Then prepare the negotiation with an advisor that never sends anything without you.

No account · no AI call · try a sample:

Sample result · fictional Series A term sheet

9 issues found, 2 standard protections missing.

  • Deal-breakerFull-ratchet anti-dilutionfull ratchet

    “Anti-dilution: In the event of a subsequent financing at a lower price, the Series A conversion price will be reduced on a full ratchet basis.”

  • MajorOption pool20% pre-money

    “Option pool: The Company will reserve a 20% option pool created pre-money, so that the pool is included in the pre-money capitalization.”

  • MajorLiquidation preference multiple2x

    “Liquidation preference: The Series A will carry a 2x liquidation preference, participating, with no cap on participation.”

See the full check →
open clause rules
48open clause rules
missing-protection checks
17missing-protection checks
document types
4document types
hard cap on every run
$10hard cap on every run
messages sent without you
0messages sent without you

Two ways in

Start with the check. Go deeper when it matters.

Free · no account

Instant contract check

Paste the document. See each clause that deviates from market norms — the exact text, why it matters, what's normal, and the redline to ask for — plus the standard protections it's missing. Rule-based, private, and done in seconds.

Workspace

Full negotiation prep

Structure the case, read the paper with the model on top of the rulebook, role-play the other side, price your walk-away, draft the reply and check it for risk. A Critic argues with every major output before you see it.

How it works

Intake → read → simulate → strategy → draft → gate.

  1. 01

    Intake

    Your goals, constraints and must-haves become a structured brief. Counterpart interests are marked as hypotheses, not facts.

  2. 02

    Read

    Attached documents are read clause by clause against the rulebook and the model — every adverse term gets a redline.

  3. 03

    Simulate

    The counterpart, role-played turn by turn with a read on every move. Illustrative ranges, never a forecast.

  4. 04

    Strategy

    Two to four approaches, each anchored on a named framework, a priced BATNA and an ordered concession ladder.

  5. 05

    Draft & check

    The message you'd send, then a risk pass for binding language, legal signals, tone and weak facts — each flag carries a fix.

  6. 06

    Gate

    Exporting the recommended message is blocked until a person approves it. Clarity never sends.

Watch it work

An agent room, not a chat box.

Agents talk to each other in typed messages — a case structured, a clause flagged, a challenge raised, a revision made — and you can read the whole conversation. Every step shows what it cost and how confident it is.

The roster

Eight agents, one typed conversation.

Intake

Structure the case

Turns a raw situation into parties, interests, a ZOPA estimate, a leverage read and measurable success metrics.

Document Analyst

Read the paper

Extracts and classifies each clause, maps it to your goals, and drafts the redline — grounded in the rulebook's verified findings.

Simulation

Role-play the other side

A turn-by-turn negotiation with a persona card, a read on each move and three scenario cards.

Strategy

Options + BATNA

Distinct approaches, each with a named framework, a priced walk-away and an ordered set of graceful concessions.

Drafting

Write what you send

The email and talking points for the chosen strategy — for a human to send.

Risk Checker

Hard stops

Over-commitment, legal signals, tone and factual weakness, plus a domain checklist. A high-severity flag forces human review.

Critic

Argue with the output

A pre-mortem on every major stage across five dimensions; every challenge carries a counter, and the Reviser folds it back in.

Outcome Tracker

Learn from the result

Records the confirmed outcome and writes reusable tactics to a case playbook you can view, edit and delete.

Grounded, not guessed

An open rulebook under the model.

Clarity doesn't just ask a model “is this contract fair?” A deterministic rulebook reads each clause, pulls out the number that matters — the pool percentage, the notice window, the uplift — and judges it against a stated market norm. Those verified findings go to the model as evidence, and if the model leaves out a deal-breaker, the rulebook puts it back.

Read all 48 rules →

It's good at

  • Reading the number in a clause
  • Not flagging things that are fine (negations, mutual clauses, IP that transfers on payment)
  • Citing the clause and the norm

Where it stops

  • Unusual wording can slip past the rules
  • It can't judge your specific leverage
  • It is not legal advice

Human authority

Clarity never sends, signs, or commits for you.

It escalates — with the reasoning trace and its confidence — whenever any of these is in play. The recommended next message cannot leave the system until a person approves it.

Language that could bind you
High-severity legal / ethical / reputational flags
Regulated domains
Low-confidence or missing facts

Why you can trust the output

Uncertain by design, costed, on the record.

Uncertain by design

Simulations are ranges with persona cards, never predictions. Confidence is shown per stage and overall.

Costed

Every run has a hard $10 cap. A live token meter shows what each agent step spent.

On the record

Typed messages between agents, a full audit log and a per-step inference trace — for every run.

Questions

Straight answers.

Is this legal advice?

No. Clarity is decision support: it tells you what to question and what to ask for, and helps you prepare. It doesn't create a lawyer-client relationship — have counsel review anything binding.

Does the instant check send my document to an AI?

No. The instant check is rule-based and runs without a model call. Your document is analysed in memory and never stored; only the short clause quotes in your result are kept, behind a private link that expires in 30 days.

How accurate is it?

When a rule fires it's usually right — it cites the exact clause and the number it read. What it can miss is unusual wording: in our tests on documents the rules had never seen, they recognised between about a third and two-thirds of the issues we planted (and were right about roughly 8 or 9 in 10 of what they flagged), which is why every result says a clean check isn't a clean bill of health, and why the full analysis adds a model read on top.

What does it cost?

Instant checks are free — 3 a day without an account. A free account adds 3 full agent runs a month; a team seat is $39 a month. Every run is capped at $10 of model spend.

Will it ever send something on my behalf?

Never. The recommended message can't be exported until a person approves it, and Clarity has no way to email, sign or submit anything.

Can I delete my data?

Yes — instantly, from your workspace, along with a full JSON export. Instant-check results can be deleted by whoever ran them and expire on their own after 30 days.

Find out what you're about to agree to.

Free, private, and done in seconds. No account needed.