Patent77

Patent77 is a patent-focused search and reporting service for inventors, small-firm patent professionals, and AI agents that need a structured view of potentially relevant patent references. Its combination of cited results, paid patentability reports, and API and MCP access makes it a practical starting point for early assessment; its search limits make it a poor substitute for a complete clearance or legal advice.

Overview

Patent77 searches the live Google Patents record, including US and major international patent collections. It turns a plain-English invention description into claim-style elements, expands synonyms, iterates queries, reads candidate documents, and scores relevant references. The result is more useful than an undifferentiated keyword list when an inventor needs an initial view of possible overlap, though the knockout-risk read should be treated as a prompt for closer review rather than a filing decision.

Publication-number lookups are never billed, a useful distinction for users who need to inspect known patents without spending search or report capacity. The search corpus does not include non-patent literature, unpublished applications within the 18-month publication window, or collections outside its coverage. That boundary matters: Patent77 can surface patent references, but cannot establish that an invention is clear across all relevant prior art.

Key features

Search and reports

Semantic search and claim analysis support the query process. Paid patentability reports add an element-by-reference map, an examiner-style novelty and obviousness narrative, limitations, and PDF export. Those structured components can help an inventor or practitioner organize a preliminary review; users who need only occasional reference searches may not need the paid report allowance.

API and MCP access

The JSON API and hosted MCP server expose search_prior_art, get_patent, and assess_patentability. This makes Patent77 a plausible fit for developers and AI agents that need patent search and lookup in a workflow, rather than only a browser-based tool. Anonymous API access allows searches and lookups without an account, but is capped at 4 requests per minute, 20 per day, and three references per search; it is a constrained way to evaluate integration, not a production allowance.

Privacy and data handling

Patent descriptions are saved within the user's account and are not used to train models, shown to other users, or sold. Descriptions are sent to an AI provider for query planning and reference reading, while derived queries—not raw descriptions—go to the public patent corpus. The service says it uses TLS, managed Postgres through Supabase with row-level security, and hashed API keys. Deleting an account removes searches, reports, monitors, and keys, with backups expiring on a fixed schedule. Users with sensitive invention disclosures should weigh the AI-provider processing alongside these stated protections.

Pricing

Patent77 is freemium, with paid plans starting at $49 a month. There is no free trial, but the free tier is available indefinitely.

The free plan keeps the cost at zero but limits both monthly search volume and references per search, and it does not include API or MCP access. Pro adds those integrations and a larger quota, but reports beyond ten cost extra. Scale raises the search, report, request, and monitor limits and adds priority support and an uptime SLA; it is hard to justify for occasional research. The published plan terms give monthly billing, but no seat allowance is stated, so the plans should not be treated as team-priced licenses.

Platforms

Patent77 is available through a web app and API; its hosted MCP server also supports agent-oriented access. The free plan is web-only, while Pro and Scale include API and MCP. That distinction makes the free tier suitable for trying the search experience, not for evaluating the paid integration workflow in sustained use.

Who it's for

Inventors and founders can use the free allowance to identify patent references before deciding whether to seek professional help. Attorneys and agents at small firms may value the claim-oriented search and structured reports as an early research aid, while AI-agent builders have a direct API and MCP route. It is less suited to work requiring non-patent literature, unpublished applications, or patent collections beyond its corpus, and its own legal limitation rules out treating a report as legal advice.

Pros and cons

Pros

• Searches start from plain-English invention descriptions and return cited references with a knockout-risk read, giving early-stage users more context than a bare lookup.
• Publication-number lookups are never billed, so checking a known patent does not consume the paid search or report budget.
• Paid reports provide an element-by-reference map, novelty and obviousness narrative, limitations, and PDF export for a more organized preliminary assessment.
• API and hosted MCP access, plus a free web tier, give both browser users and developers a way into the product.

Cons

• The corpus excludes non-patent literature, unpublished applications in the 18-month publication window, and collections outside its scope, limiting its value for exhaustive prior-art work.
• The free tier allows just 3 searches monthly and returns only the top 5 references, which can constrain repeated exploration.
• Pro charges $4 for each report beyond its monthly ten, while Scale costs $199 a month; users with sporadic needs may pay for capacity they do not use.
• Search descriptions are processed by an AI provider, a consideration for users handling confidential invention details despite the stated account and security protections.

Alternatives

For a broader comparison, browse AI Patent Search Software.

• FindIP is worth considering when mobile access or a wider free search allowance matters: it supports Android, iOS, API, and web, and its free plan includes 30 search credits, all APIs, five jurisdictions, and re-ranking.
• GoVeda may suit users prioritizing broad patent volume and generous free semantic search: its OpenPatent plan is free forever and offers 220M+ patents worldwide and unlimited AI semantic searches, as well as basic metadata filters and AI reports and summaries.
• Patently is an option for web users seeking asset search with filters, patent-family browsing, and chat with Onardo; its free plan is limited to one user and individual use.
• PatLab.AI is another web-based patent-search option, with free searches and a one-time 49.00 USD Pro Patent Search plan.
• PatSnap is worth comparing for users who want technical-research capabilities beyond patent search: its free Basic plan includes limited technical topics, basic AI technical Q&A, TRIZ innovation generation, and report previews; it also supports API, self-hosted, and web access.
• PQAI may appeal to developers who favor an open-source platform; it offers API and web access and has a free plan.
• The Lens offers anonymous free browsing and a free academic workspace, making it an alternative for users who want those access options.
• Stiltn offers a full-access evaluation with dedicated onboarding and custom configuration, and may suit users who want to assess a self-hosted or web platform.

Verdict

Patent77 is best for inventors, small-firm patent professionals, and AI-agent builders who want a claim-oriented patent search with cited references and an optional structured report. Its free tier and unbilled publication-number lookups lower the barrier to preliminary research, while Pro adds useful API and MCP access. Choose another route or consult a registered patent attorney or agent when the decision requires broader prior-art coverage or legal judgment.