Sparlo turns an invention disclosure into a defensible patent moat — every application worth filing, the prior art around each, the claims that survive a challenge.
Built for privileged work: the disclosure is never sent to outside search, never used to train a model, and locked to your account — so a live, unpublished invention can’t become prior art against your client. Read the security review
Invention BriefInvention: a phase-change thermal interface for high-power EV battery modules — a eutectic gel channel routed between the bipolar plates that holds peak cell temperature under 45 °C at sustained 4C discharge. Novel elements: the gel’s eutectic composition, the serpentine channel geometry, and the plate-integrated fill ports. Known prior art: forced-air cooling, liquid cold plates, immersion baths.
A complete patent-moat report — the invention generalized to its core principle, a map of every application it can claim, the prior art that surrounds each one, the claim architecture where the durable protection lives, and a filing strategy — then an adversarial stress test attacks all of it before an examiner can.
Each disclosure runs a privileged eight-stage pipeline that ends by attacking its own work — every prior-art claim grounded with citations, every inventor premise audited against the art.
Generalize
Abstracts the disclosure to its governing principle, so the moat is built around the invention — not one narrow embodiment — and logs every premise the inventor asserts as fact, for the stress test to audit later.
Application Map
Maps every field the invention could claim — including the far-field, non-obvious domains a practitioner in the origin field would never list. The cross-domain surface most disclosures leave on the table.
Prior-Art Differentiation
Searches patents and technical literature for each application and differentiates element by element — at claim level when claim text is retrievable, and labeled honestly when it is not.
Inventive Step
Builds the §103 argument under Graham and KSR — the articulated reason to combine, the reasonable expectation of success, and secondary considerations checked for the nexus each one needs.
Claim Architecture
Locates the durable protection: the broadest mechanism the disclosure supports, organized into statutory claim sets — each with its closest art and the distinguishing limitation that survives it.
Claim Scaffold
Plain-language claim concepts you finalize — independent mechanism claims, a dependent ladder of fallback positions, and blocking claims for the obvious design-arounds. Never drafted claim language.
Stress Test
Turns adversary: builds the strongest §102/§103 case against each limitation, audits every asserted premise against the art, then runs its own refutation searches and finalizes against what comes back.
Filing Strategy
A jumbo provisional with a shared detailed description, converted into US and foreign filings with differentiated claim sets — plus the invention-split call and the disclosure gaps each filing still needs.
Map the full moat
The cross-domain application map — every adjacent field the invention can claim, not just the one in front of you.
Invention: a eutectic phase-change gel channel routed between EV battery bipolar plates, holding peak cell temperature under 45 °C at sustained 4C discharge.
Differentiate from prior art
A grounded prior-art search for each application, with the distinguishing features articulated against the closest references — by citation, with one-click verify links.
Distinguish the plate-integrated eutectic geometry from liquid cold plates, immersion cooling, and air cooling for high-power battery packs.
Stress-test before the examiner does
An adversarial pass builds the strongest §102/§103 case against each claim limitation, audits the premises the inventor asserted as fact, and runs its own refutation searches — so the weakness surfaces now, not in prosecution.
Stress-test the inventive-step story for the thermal-interface invention against the retrieved art — run the searches a skeptical examiner would.
Plan the filing
A jumbo provisional with one shared detailed description, converted into US and foreign filings with different claim sets — fewer attorney hours, broader coverage.
Plan the filing for the thermal-interface invention: provisional scope, then US and foreign conversions across the mapped applications.
Why not just ask ChatGPT?
A general AI chat is built to sound convincing. This is built for patent diligence.
General AI chat tools (ChatGPT, Claude) are capable — but they answer in prose, cite from memory, and run on consumer terms. A patent moat needs structure, grounded citations, and a disclosure that never leaves your control. Here is the difference, dimension by dimension.
Built for §102 diligence
A general-purpose chatbot
A privileged eight-stage pipeline built for §102/§103 diligence — application map, prior-art differentiation, claim architecture, and an adversarial stress test.
A general-purpose chat that answers whatever you ask. It was not built for patent diligence, and the shape of its answer is a conversation, not a moat structure.
A full-recall prior-art search for each application, run against patents and technical literature — outbound queries reach Exa and Google Patents, never general web search.
A single-pass answer from what the model already holds, or one round of browsing. No structured search across the patent and non-patent literature for each application.
Differentiates element by element against the closest references — at claim level when the claim text is retrievable, and labeled honestly when it is not.
A prose summary of what it recalls about the field. No element-by-element mapping to the closest art, and no distinction between claim text it retrieved and text it inferred.
Grounded prior-art citations with one-click verify links — it cites only documents it actually retrieved, so every reference is one you can pull and check.
Citations recalled from training, not retrieved — which is exactly where fabricated case and patent cites come from. You have to verify every one before you can rely on it.
An adversarial stress test builds the strongest §102/§103 case against each limitation, audits every asserted premise against the art, and runs its own refutation searches.
It will critique on request, but it does not turn adversary against its own output, audit the premises, or run fresh refutation searches to try to break the position.
The disclosure never goes to outside search: the refiner has no web access, and the disclosure document is never sent to any search provider. Encrypted in transit and at rest, locked to your own account at the database layer — never used to train AI models, at Sparlo or at Anthropic.
Consumer AI chat tools keep your prompts on servers you do not control, on default terms that can train on what you enter. Pasting an unpublished disclosure there is the exposure a §102 review exists to prevent.
A finished analysis you refine, revisit, and share on a passcode-gated link — the shared page serves the finished analysis only, never the disclosure text.
A chat thread that scrolls away. No structured report to reuse across a matter, and no passcode-gated way to share the work product without exposing what you typed in.
Run one disclosure through the pipeline and see the difference for yourself — your first prior-art search is included.
Start free — first search includedExample Library
See it on a real patent.
Every analysis below ran on a real, public patent — the same pipeline your disclosure would. Pick one and see the actual work-product: the applications it maps, the prior art it pulls, the claims it stress-tests.
Patent Moat Analysis · Biotech
Nucleic-Acid Sequence Amplification by Primer Extension
Prior-art & non-obviousness analysis — nucleic-acid amplification
8-Section deliverable
Retrieved prior art · sample
- US20220033891A1Amplification with primers of limited nucleotide composition
- JP2004526442A5
- US11466315B2Fast PCR for STR genotyping
- CN114250278BMethods, compositions and kits for capturing, detecting and quantifying small …
- CN116121349BA method for amplifying a nucleic acid sample containing ethanol
+ 23 more references retrieved and screened
Confidential by construction — built for §102.
The disclosure never goes to outside search
The refiner has no web access, and the disclosure document is never sent to any search provider — outbound prior-art queries reach only Exa and Google Patents, never general web search.
Encrypted, isolated, never trained on
Encrypted in transit and at rest, locked to your own account at the database layer — never visible to your team, never used to train AI models, at Sparlo or at Anthropic.
Deleted on demand
Delete a report and the disclosure text and any uploaded file are removed from storage immediately — not on a schedule.
Built to survive your firm’s security review.
In patent work a careless disclosure isn’t a privacy incident — it’s prior art against your own client. Below are direct answers to the questions a Rule 1.6 diligence review will ask, stated specifically enough to be checked.
The path a disclosure takes
Your browser → Sparlo
The disclosure travels over TLS to Sparlo and is stored in your private account in the United States (US-West). Row-level security at the database layer — not application code — means only your login can read it.
Locked to you alone
IP analyses are pinned to your personal account by a database trigger. They are never visible to teammates and never co-mingled with other firms. Nothing leaves your account unless you explicitly create a share link — and a shared page is passcode-gated and serves the finished analysis only, never the disclosure text.
Analysis on the Anthropic API
Model calls go directly to Anthropic under its Commercial Terms: no training on customer content, automatic deletion of API inputs and outputs within 30 days. Provider failover is disabled for IP analyses — your content is never rerouted to an alternate AI provider.
The report — and the delete button
The finished analysis lives in your account until you remove it. Deleting is immediate and hard: the disclosure text, the analysis, and any uploaded file are purged from the database and file storage the moment you click — not on a schedule.
What your GC will ask
Is our client’s disclosure used to train AI models?
No — at both layers. Sparlo does not train or fine-tune any model on customer content. Model calls run on the Anthropic API under Anthropic’s Commercial Terms of Service, which state that Anthropic "may not train models on Customer Content from Services." Anthropic also automatically deletes API inputs and outputs within 30 days.
Does the disclosure get sent to a search engine?
The disclosure document is never sent to any search provider. Outbound prior-art queries are restricted to two providers — Exa (technical literature) and Google Patents — and general web search (Perplexity, Tavily) is disabled for IP analyses in code. During the pilot, analyses run in full-recall mode: queries are built from generalized invention terms so prior-art recall — finding the art that actually matters — is maximized. We state that plainly rather than bury it. A stricter scrubbed mode also exists in code — queries restricted to search-safe framings, with the protected terms you confirm blocked by a deny-list gate before anything leaves — and your firm can elect it instead. The trade-off is yours to make.
Could it leak through a share link?
Not unless you create one — and even then, two factors stand between the link and the content. No share link exists until you, the account owner, explicitly generate one. Links for IP analyses use unguessable 122-bit tokens plus a separate passcode: the page serves nothing until the passcode is entered, it is meant to travel separately from the link, and Sparlo stores only a salted hash of it — never the passcode itself. Entry is attempt-limited, links expire after 30 days, regenerating rotates the passcode, and revocation is immediate. Even unlocked, a shared page renders the finished analysis only; the underlying disclosure text is never served through the public path.
Is it encrypted?
In transit, TLS for every connection. At rest, AES-256 at the storage layer. Background-job payloads are additionally application-layer encrypted and carry only record IDs — the disclosure text itself never enters the job queue or its logs.
How long do you keep it — and who controls that?
You do. The analysis stays in your account so your work product persists; deletion is user-controlled, immediate, and hard — database row, disclosure text, and uploaded files are purged at once. Upstream, Anthropic deletes API inputs and outputs within 30 days of processing (except where law requires retention or content is flagged under its usage policy).
Which model is this? Are there hidden subprocessors?
Claude, by Anthropic, called directly — no resold model marketplace, no failover chain. Anthropic holds SOC 2 Type II, ISO 27001:2022, and ISO/IEC 42001:2023 certifications, and publishes its trust documentation and subprocessor list at trust.anthropic.com.
For your ethics review
Rule 1.6 — confidentiality
ABA Formal Op. 512 makes "how does this tool handle client data" a mandatory diligence question. This page is the answer: no training, restricted egress, account isolation, user-controlled deletion — stated specifically so your ethics review can verify rather than trust.
Op. 512 — no self-learning system
Opinion 512 attaches an informed-consent burden to self-learning generative AI that retains and reuses prompts across matters. Sparlo is not a self-learning system: your inputs do not train models or carry over into anyone else’s analysis.
37 CFR 11.18 — you remain the practitioner
Sparlo is a co-pilot, not an autopilot. Findings cite verifiable sources — patents and literature you can pull and check — because USPTO guidance is explicit that relying on a tool’s accuracy does not satisfy your reasonable-inquiry obligation. The product is built around your review, not in place of it.
Duty of candor — no new disclosure burden
Under the USPTO’s April 2024 guidance, there is no general duty to disclose AI assistance in a submission unless it is material to patentability. Sparlo produces analysis and strategy support — it draws no legal conclusions, and the judgment calls stay yours.
Sources your reviewer can verify: Anthropic Commercial Terms of Service and certifications (trust.anthropic.com) · ABA Formal Opinion 512 (July 2024) · USPTO, Guidance on Use of AI-Based Tools in Practice, 89 FR 25609 (April 2024). Questions this page doesn’t answer — ask us directly and we’ll answer in writing.
Pricing
Priced per prior-art search, not per seat.
Start free — your first prior-art search is included, no credit card required. Every plan runs the same complete moat analysis.
billed monthly
4 prior-art searches / mo
For the solo practitioner running the occasional moat analysis.
Start freebilled monthly
10 prior-art searches / mo
For attorneys building patent moats as a regular part of practice.
Start freebilled monthly
20 prior-art searches / mo
For teams with steady filing volume across matters.
Start freeEvery plan includes
- The full §102/§103 moat report — application map, prior-art differentiation, claim architecture, and an adversarial stress test
- Grounded prior-art citations with one-click verify links
- The disclosure never leaves for outside search and is never used to train a model
- Passcode-gated share links — the shared page serves the finished analysis only, never the disclosure text
- Locked to your account, deleted on demand
Build the moat before you file.