Per Seat, Per Matter, Per Gigabyte: How Legal AI Pricing Models Actually Work

Two legal AI tools can appear similar in price yet cost wildly different amounts at your actual caseload. For example, a $300 flat monthly fee and a $300 per-demand fee are the same number. They are not the same bill. At your actual matter volume, one will cost a fraction of the other. The challenge is that legal AI vendors do not structure their pricing on the same axis. One tool charges by the user, another charges by the document, and a third charges by the gigabyte.

We track 43 legal AI tools. Across these platforms, we have identified seven distinct pricing models currently in use. Understanding these models is the only way to compare different quotes side by side. If you are still trying to secure a base quote, our guide on Why So Many Legal AI Vendors Hide Their Pricing (And How to Get a Real Number) covers how to navigate sales calls. Once you have those numbers, this post serves as your decoder.

Some of these platforms are pay-as-you-go, while others require annual commitments. Our analysis of Why Almost No Legal AI Tool Offers a Free Trial (And How to Evaluate the Ones That Don't) explains why trial policies vary so much. Below, we break down each model, explain who they favor, highlight hidden gotchas, and provide a framework to project what you will actually pay over a year.

The pricing models, explained

1. Per-seat / per-user subscription

How it works: Your firm pays a fixed monthly or annual rate multiplied by the number of licensed users. The price is predictable and scales with headcount, not with your actual output.

Who it favors: This model is best for firms where multiple attorneys will use the tool daily. It is also highly favorable for solo practices, because a single seat represents the lowest possible starting cost for that software.

Hidden gotchas: As your firm grows, per-seat costs compound. The total cost can quickly surpass what a flat-rate or output-based system would charge. Monthly billing is typically 10 to 15 percent higher than annual commitments. This model also creates internal politics. Partners must decide who receives an AI license, which often leads to shadow usage and security risks when unlicensed staff share credentials.

Examples: Clio Manage AI is bundled into the Clio Manage Complete plan starting at $149 per user per month when billed annually, according to Clio's published rates. Lower tiers do not include the full AI features.

Pros

• AI is embedded directly in the practice management platform most solo and small firms already use.
• Automates deadline extraction from court documents and drafts emails or letters from case data.

Cons

• AI features require the Complete plan, which is a higher entry cost for solo practitioners on a tight budget.
• AI only accesses data inside Clio and cannot search external case law or statutes.

Paxton AI costs $499 per user per month on a month-to-month basis, as shown on Paxton AI's pricing schedule. An annual commitment reduces the effective cost to $250 per user per month, billed at $2,999 per user per year.

Pros

• Purpose-built family-law workflow that reviews intake forms, financial disclosures, and custody evaluations.
• Generates parenting plans and settlement agreements.

Cons

• The month-to-month rate of $499 is expensive for solo practices.
• Generalist platform where family law is only one vertical among many.

2. Per-matter / per-demand (output-based)

How it works: Your firm pays a flat fee each time the software generates a specific output. There are usually no seat limits and no monthly minimums.

Who it favors: This model favors firms with variable or highly cyclical caseloads. It is also ideal for practices where each matter generates a discrete, high-value deliverable, such as a demand letter, which directly recoups the software cost.

Hidden gotchas: At sustained high volumes, per-demand pricing compounds quickly. Add-ons and page limits can push the actual cost far above the advertised base rate. A solo attorney processing 5 demands a month pays $1,500. At 50 demands a month, that bill climbs to $15,000. A per-seat subscription at $250 per month would easily outperform this model at higher volumes.

Examples: EvenUp charges a starting base rate of $300 per demand, but competitor analysis shows that add-ons can push the effective cost to $500 to $800 per demand. The company does not publish its pricing publicly.

Pros

• Purpose-built for personal injury demand letters.
• Combines AI drafting with in-house legal reviewers for quality control.

Cons

• Per-demand pricing is expensive at high case volumes.
• Pricing is not published publicly and requires a sales call.

Precedent charges a flat $275 per demand, which includes unlimited pages and unlimited revisions, according to Precedent's public comparison.

Pros

• Published flat-rate pricing with no hidden add-ons or page token charges.
• Includes a full personal injury lifecycle coverage and an automated Exhibit Manager.

Cons

• No public funding announced, making it a smaller vendor compared to heavily funded competitors.
• Specialist tool with no utility for non-personal injury practice areas.

3. Per-gigabyte data hosted (ediscovery)

How it works: Pricing scales with the volume of data stored on the platform, measured in gigabytes after processing and deduplication. Core AI features, such as document Q&A and deposition analysis, are often bundled into the base hosting rate.

Who it favors: Litigation teams managing large discovery matters benefit from this model. Headcount does not affect the price, so you can have 10 attorneys reviewing a single 500-gigabyte data set without paying for 10 separate licenses.

Hidden gotchas: Raw files often expand significantly after processing and extraction, meaning your billable gigabytes will be higher than the size of the original upload. Some platforms charge one-time ingestion fees when you first upload data. Heavy batch operations, like running AI summaries across millions of documents at once, may require you to purchase additional compute credits.

Examples: Everlaw uses per-gigabyte hosted pricing. It bundles its AI Review Assistant, Deposition Analyzer, and Writing Assistant into the base rate, but its Deep Dive tool requires an extra, one-time per-gigabyte ingestion fee. Everlaw does not publish its rates, but third-party estimates for mid-market matters run between $2,000 and $5,000 per month.

Pros

• Deposition Analyzer AI and single-document Q&A are included at no extra cost.
• Cloud-native architecture with predictable hosting rates.

Cons

• Estimates of $2,000 to $5,000 per month put it out of reach for small firms without large discovery matters.
• Batch AI operations require credits beyond the base rate.

DISCO bills per gigabyte of processed data. Its Cecilia AI tool is bundled with no ingestion fees under its 2025 all-inclusive platform. DISCO does not publish its rates publicly.

Pros

• Cecilia AI is included in the base per-gigabyte price for fact queries and deposition summaries.
• All-inclusive platform covers eDiscovery, deposition management, and timelines in one product.

Cons

• Enterprise licensing is prohibitive for small plaintiff boutiques.
• Handling is restricted to litigation review and does not support external legal research.

4. Pay-per-export / pure usage

How it works: The platform is free or very cheap to access. You only pay a fee when you export a finished document, abstract, or report.

Who it favors: This model is ideal for firms with occasional, project-based needs. If you only review a specific type of contract a few times a year, you pay nothing during the months you do not use the tool. There are no upfront or ongoing subscription commitments.

Hidden gotchas: If your project volume increases, pay-per-export costs escalate rapidly. Because there is no ongoing subscription, you do not receive volume discounts, dedicated support, or guaranteed processing capacity. Firms that begin using these tools daily will quickly outgrow them on cost alone.

Examples: LeaseLens charges a flat $25 per lease export. It offers a free tier to access the platform, as detailed on the LeaseLens pricing details page.

Pros

• Free tier available with no subscription required.
• Low barrier to entry for transactional firms with variable volume.

Cons

• Narrow feature set limited to lease abstraction only.
• Pay-per-export model becomes expensive at scale, such as 100 leases costing $2,500.

5. Credits / usage-based overages

How it works: You pay a base monthly subscription to access the platform, which includes a set amount of usage. If you perform resource-intensive tasks, such as transcription, optical character recognition (OCR), or complex AI queries, you consume credits. Once you exhaust your monthly credit allotment, you pay overage fees.

Who it favors: This model suits firms with steady baseline work but unpredictable monthly spikes. The base subscription stays affordable, and the credit system handles the extra computing costs during busy months.

Hidden gotchas: Credit overages are highly unpredictable. If you do not actively track page counts, transcription minutes, and AI queries, your monthly invoice can easily double or triple without warning. This makes annual budgeting extremely difficult.

Examples: Casefleet offers a Starter plan at $30 per month and an Advanced AI plan at $140 per month, according to Casefleet's published pricing schedule.

Pros

• AI Document Intelligence auto-extracts, summarizes, and links facts across discovery.
• Audio and video reviewer handles body-cam footage and surveillance video natively.

Cons

• AI features are gated at the $140 Advanced tier, while the $30 Starter tier has limited AI capability.
• Usage-based overages on transcription can add up quickly on high-volume audio discovery.

ProPlaintiff.ai uses a four-tier subscription system. Additional credits cost $0.10 each, and seat add-ons cost between $49.99 and $89.99 per seat. Its base subscription pricing is not publicly published.

Pros

• Published pricing page adds buyer confidence.
• Credit-based model allows pay-as-you-go without annual commitment.

Cons

• Credit system complexity makes monthly costs hard to forecast.
• Newer brand with less customer proof compared to larger competitors.

6. Flat per-firm subscription (not per-seat)

How it works: Your firm pays a single monthly fee for the software. This fee covers your entire firm, regardless of how many attorneys or staff members use the platform.

Who it favors: Small firms where multiple team members need shared access to the same tool. For a three-attorney firm, this removes the need to buy three separate licenses.

Hidden gotchas: Flat pricing is rare in the legal AI space. Vendors often use this model to gain early market share. Once the tool becomes essential to your daily workflow, the vendor may introduce per-seat tiers at contract renewal.

Examples: Briefpoint charges a flat $89 per month. It features a single tier with no per-user fees and no enterprise upsells. It is the only tool we track with a published, single-tier flat-firm price.

Pros

• Single flat price with no per-user fees is highly accessible for small practices.
• Drafts discovery requests and responses in minutes.

Cons

• Extremely narrow scope that does not draft motions, research case law, or manage practice.
• No audio or video review capability.

7. Bundled add-on to a larger suite

How it works: AI features are not sold as a separate, standalone product. Instead, they are included in high-tier subscriptions of your existing practice management or legal research software.

Who it favors: This model is best for firms already deeply committed to a specific software ecosystem. If you already use the parent platform, you do not have to manage another vendor or integrate new software.

Hidden gotchas: The AI is restricted to the data stored within that specific platform. A practice management AI cannot research case law. A research AI cannot access your firm's billing logs. Furthermore, the true cost is the total subscription price, not just the AI features. You must pay for the entire high-tier platform to access the AI.

Examples: MyCase IQ is included in the MyCase Pro plan, which costs $100 per user per month billed annually.

Pros

• AI is embedded in MyCase practice management, requiring no separate tool or login.
• Automated time-entry generation reduces billing leakage for solo practices.

Cons

• AI accesses only MyCase workspace data and cannot perform external legal research.
• Smaller integration ecosystem compared to larger practice management platforms.

Lexis+ with Protege is an add-on to a Lexis+ research subscription. The AI add-on costs between $125 and $275 per user per month. When combined with the base Lexis+ subscription, the total cost for a solo practitioner can reach $300 to $675 per month.

Pros

• Combines AI drafting assistant with full case law and statute databases.
• Stanford study found a 17 percent hallucination rate, which is half that of competing research platforms.

Cons

• Pricing requires a sales conversation with no self-serve rates.
• High total cost is materially more expensive than practice-management-bundled AI options.

At a glance

How to compute your true annual cost

1. Estimate your monthly volume on the relevant axis. Count how many seats you need, how many demands you generate, or how many gigabytes of data you host.
2. Multiply your estimated volume by the unit price to find your base monthly cost.
3. Identify potential overage triggers. Read the fine print for page limits, transcription minutes, and extra credit costs. Estimate how often you will exceed these limits.
4. Factor in implementation or onboarding fees. These are especially common in enterprise ediscovery and bundled contracts.
5. Multiply your monthly total by 12 to annualize the cost. If you are comparing annual commitments to monthly plans, note that annual billing usually lowers the effective rate, and at some vendors the cut is steep: Paxton AI's annual plan, for example, drops its effective rate from $499 to about $250 per user per month.
6. Add the opportunity cost of switching. Many tools require 90-day minimums or annual commitments. Budget for the transition period if you are running two tools at once.

Worked example: break-even between per-export and per-seat

A solo real estate attorney uses LeaseLens for lease abstractions. At the current $25 per export rate:

• 5 exports per month equals $125 per month, which is $1,500 per year.
• 10 exports per month equals $250 per month, which is $3,000 per year.
• 20 exports per month equals $500 per month, which is $6,000 per year.

Now, compare this to Clio Manage AI. The Complete plan, which includes these AI features, costs $149 per user per month billed annually. For a solo attorney, this is $1,788 per year.

Clio Manage AI provides general drafting, calendar automation, and time entry across your entire practice. LeaseLens is a specialized tool built only for lease abstraction. While they do different things, this comparison illustrates the core pricing dynamic. Per-export pricing is cheapest at low, irregular volumes. It becomes the most expensive option once the work becomes a daily habit.

At 10 exports per month, LeaseLens alone costs $3,000 per year, which is more than the full Clio Manage AI suite. If your primary volume is in the personal injury or per-demand space, our guide on pricing transparency covers how to run similar break-even math at 50 cases per month.

FAQ

Which legal AI pricing model is cheapest for a small firm?

No single model is cheapest in the abstract. It depends entirely on your specific caseload. Flat per-firm models, such as Briefpoint at $89 per month, are highly affordable if the features fit your workflow because they ignore headcount. Pay-per-export models, such as LeaseLens at $25 per export, are cheapest if your volume is extremely low. Per-seat models become cost-effective when your usage is daily and your headcount is small. If you handle fewer than 5 matters per month that require AI, per-demand or per-export models usually win. Above that volume, per-seat or flat-firm subscriptions typically cost less than per-output platforms.

Per-matter or per-seat: which is better at low volume?

Per-matter pricing favors low-volume firms because you pay nothing during quiet months. A solo attorney who generates 3 demand letters per month pays $825 with Precedent. A per-seat subscription might cost between $250 and $499 per user per month regardless of your output. The per-seat option only wins once your output is frequent enough that the flat monthly subscription is cheaper than the accumulated per-unit fees. Calculate your average monthly output over the last 12 months to find your break-even point.

Why do ediscovery tools charge per gigabyte?

EDiscovery platforms must process, index, and host massive volumes of files. Their infrastructure costs scale directly with data storage, not with headcount. Passing this variable cost to you ensures that firms only pay for the data footprints they actually use. Everlaw and DISCO both use per-gigabyte models but handle ingestion fees differently. Everlaw charges a one-time per-gigabyte fee for its Deep Dive feature at ingest. DISCO eliminated ingestion fees entirely on its 2025 all-inclusive platform. This structural difference can change your initial upload cost, even if the ongoing hosting rates are similar.

How do I compare two tools priced on different models?

Convert both tools to a projected 12-month total based on your firm's volume. Estimate your required units, whether they are seats, outputs, or gigabytes. Multiply by the unit price, add potential overages or onboarding fees, and annualize the number. Once you have a single annual figure, ask if the price includes the specific features you need, if there is a minimum volume commitment, and what triggers a price increase at renewal. Expressing both options as a single annual dollar figure under the same assumptions allows you to compare them fairly.

The Bottom Line

Legal AI is sold under seven distinct pricing models, each reflecting a different cost driver. Your headcount, output volume, data storage, or firm size will determine your ultimate cost. The model that is cheapest for one practice can easily become the most expensive for another. There is no universal winner.

Let your volume guide your decision. Low, irregular usage favors per-output or pay-as-you-go models. Daily, multi-user work favors per-seat or flat-firm subscriptions. Large litigation data sets will force you into ediscovery per-gigabyte pricing, regardless of your personal preference. Bundled AI makes the most sense if you are already using the parent practice management platform and the features are included in your tier.

Always run the annualization math before signing a contract. Annual commitments can cut your effective monthly rate substantially, but they remove your flexibility to switch tools. Use our calculation framework to project your actual usage over 12 months. If you are still waiting on exact quotes from a sales team, read our guide on pricing transparency to help you secure those baseline numbers.