Patlytics

What it is

Patlytics is an AI platform for the whole patent lifecycle: invention disclosure, application drafting, office-action analysis, patent and non-patent-literature search, claim construction, invalidity analysis, freedom to operate, SEP analysis, portfolio pruning, and infringement detection. Paul Lee and Arthur Jen founded it in 2024. The product surface is a set of modules plus Patlytics Agent, one chat interface the vendor describes as running on more than 150 Skills.

Where it actually differs from the rest of the AI patent field is the center of gravity. Most of the funded entrants started at drafting — the attorney writing claims in Word. Patlytics started at the litigation and licensing end, where the deliverable is a claim chart backed by evidence of use, and worked backwards into prosecution. If Anaqua is the system of record for your portfolio, Patlytics is the analysis layer that reads it.

Funding runs $14M Series A in February 2025 led by Next47, then $40M Series B announced 8 April 2026 led by SignalFire with Relativity among the participants, for roughly $65 million total by the vendor’s own count. Patlytics also claims 40% of the AmLaw 100 on the platform — naming Quinn Emanuel, Latham & Watkins, Foley & Lardner, and Susman Godfrey — plus corporate customers including Rivian, Panasonic, Canon, Sanofi, and Asahi Kasei. Those are vendor figures with no independent audit behind them. Read them as scale signals, not measurements.

Why it shows up in Legal Ops stacks

• The claim chart is the product, not the demo. Detection maps a patent against a product limitation by limitation, segments long limitations automatically so evidence attaches to the right phrase, and returns an editable chart with citation-backed snippets and color-coded read-strength ratings on each mapping. Portfolio screening runs the same pass across hundreds of patents against multiple products at once.
• The MCP server is real and precisely scoped. https://mcp.patlytics.ai/mcp speaks Streamable HTTP, authenticates against a Patlytics account, and exposes five read-only tools: find similar patents, get patent details, search non-patent literature, list your portfolios, and list patents in your vault. In Claude Code that is one command:

claude mcp add --transport http patlytics https://mcp.patlytics.ai/mcp

Prior-art lookup then happens where the associate is already working. The boundary is deliberate — infringement, invalidity, FTO, and drafting stay in the web app.

• The security posture clears IP-department procurement without a side negotiation. SOC 2 Type 2, ISO 27001, ISO 42001, and GDPR, with customer data segregated, encrypted in transit and at rest, and not used to train models. History is retained 90 days and deleted on request. Assessments come from NCC Group and A-LIGN. ISO 42001 is the one to point at when the AI-governance questionnaire arrives.
• Disclosure intake is self-serve for engineers. R&D submits inventions through the platform instead of a form that lands in a paralegal’s inbox, which is where most corporate disclosure pipelines stall.

Pricing reality

Patlytics publishes no price, no tier list, and no free tier. Every deal is quoted after a demo. Third-party aggregators put it at roughly $800–2,000 per user per month, but that is an estimate extrapolated from adjacent platforms rather than a vendor or customer number — do not build a budget on it.

The anchors that are real come from the segment. Spellbook’s 2026 legal-AI pricing survey puts firm-side enterprise tools at about $1,000–2,000+ per user per month at the Harvey end, $300–800 at Legora, and $225–650 for research AI like CoCounsel and Lexis+ Protégé. Patlytics sells to that same buyer and quotes in that neighborhood.

Two things move the number, and only one of them is seats. Analysis volume is the other: detection runs and portfolio screens are the expensive unit, and a licensing program that charts 200 patents a quarter consumes a different contract than four litigators doing invalidity work. The vendor’s own ROI framing is that claim charts which ran $30K+ in attorney time now cost a fraction of that. Test it against your own number — price one matter you have already charted by hand, then compare.

Get three things into the order form: the analysis-run entitlement, the overage rate once you exceed it, and whether MCP access is included with every seat or sold separately.

Best for

In-house IP counsel and law-firm patent litigators running evidence-of-use, invalidity, and FTO work across a portfolio of 50 or more active patents, where claim charting is a recurring cost center rather than an occasional project — and where an IPMS already handles docketing and annuities.

Not for you if you file a handful of applications a year and your bottleneck is drafting speed, not charting volume: DeepIP and Solve Intelligence sit closer to that work and quote lower. Not for you if what you need is docketing, renewals, or IP spend management — Patlytics analyzes the portfolio, it does not administer it, and Anaqua stays on the invoice. And not for you if procurement requires published pricing before a demo, because there isn’t any.

Alternatives — and when to pick them instead

• Solve Intelligence — the closest full-lifecycle rival and the best-funded peer, browser-based, $40M Series B in December 2025 with M12 and Thomson Reuters Ventures participating for $55M total, and claim charting added that same month. Pick it when drafting and prosecution volume dominates your workload and charting is the smaller half.
• DeepIP — the fastest-growing entrant on the prosecution side: $15M Series A in March 2025 led by Resonance, $25M more in March 2026, and 400+ law firms and corporate IP teams. It drafts inside Microsoft Word and connects to IPMS and docketing tools. Pick it when your attorneys will not leave Word, or when chemistry and life-science drafting is the domain.
• PatSnap, Clarivate Derwent, and Questel — the incumbent patent data and analytics layer, and the two that will already be on your shortlist by market share. Pick one of these when the job is landscape analysis, search coverage, and a defensible data license rather than litigation work product.
• Harvey or Legora — generalist legal AI the firm probably already licenses. Pick it when patent matters are one slice of a mixed docket and consolidating vendors beats best fit on any single workflow.
• XLSCOUT — the budget option for teams that need prior art and an occasional chart, not a litigation platform.

Watch-outs

• Every proof point on the page is vendor-supplied. The AmLaw-100 share, the 100-hours-to-20-hours portfolio mapping, the “up to 90% reduction in project time” — all Patlytics marketing, none independently audited. Guard: make the POC chart one matter you have already charted manually, then diff it limitation by limitation. A vendor-run demo on a vendor-chosen patent proves nothing about your art unit.
• Confidence indicators are not verification. Read-strength ratings and citation links cut checking time; they do not remove attorney review, and a chart served on a bad mapping is a Rule 11 exposure that lands on a named signer, not on the tool. Guard: keep an attorney sign-off step in the workflow and log which limitations were AI-mapped, so the review record survives a challenge.
• The MCP server is read-only and account-scoped. Anyone designing a Claude Code workflow that expects to trigger infringement or invalidity runs over MCP will hit a wall — those five tools search and retrieve, nothing more. Guard: scope automation to prior-art lookup and vault queries, and keep the analysis runs in the app where the audit trail lives.
• 90-day history retention is a default setting, not a contractual term. Guard: move retention period, deletion SLA, and the subprocessor list into the DPA before signing. A trust-page statement can change without notice; a DPA cannot.
• Two years old, and the product surface moves fast. “150+ Skills” is a number that was different last quarter and will be different next quarter. Guard: pin your entitlements to a named current release in the order form, get a support horizon in writing, and refuse roadmap language — the shared firm-to-corporate workspace and the autonomous drafting agents announced with the Series B are not shipped capability you can buy today.