Liminary
Liminary is a browser-extension research memory layer that captures what you read and fact-checks deliverables inside Notion and Gmail.
Liminary is one of the few AI memory tools that treats provenance as the product rather than a footnote — every answer flags whether it came from your data or the external web, and every insight links to its source. The May 2026 Notion fact-check and June 2026 Gmail drafting move it past passive storage, and the August 2026 LongMemEval, Memora, and BEAM benchmark write-up is unusually public for this category. Pro at $29/mo is easy math if you bill hourly and lose the first hour of an engagement re-finding your own work. Skip it if you need mobile, offline, or on-prem, because it is a Chrome extension at heart. If you mainly want a general chatbot, run ChatGPT or Claude directly instead.
Verified 5d ago · liveness 80/100 · cite: rightaichoice.com/tools/liminary
- Management consultants producing client deliverables across multiple sources
- Strategy and freelance consultants advising several clients at once
- Market researchers tracking industry trends, whitepapers, and reports
- Think tank and policy analysts managing high reading volumes
- Casual note-takers who prefer manual organization in Obsidian or Notion
- Anyone needing offline or native mobile access — the core is a Chrome extension
- Teams requiring on-premise deployment or strict data sovereignty
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Liminary if you need offline access, a native mobile app, or on-premise deployment — the product is a Chrome extension, and none of those are available regardless of tier.
The free Explorer tier's 3 composer uses and 10 daily uploads run out fast, so a genuine trial quickly turns into a paid Pro decision at $29/mo.
Explorer ($0) is a real test of the capture-and-recall loop but not a working setup. Pro at $29/mo is priced for a solo consultant or analyst who bills hourly and gets the money back the first time it saves a re-research hour. Team at $50/user/mo only pays off when the shared knowledge base and 50-seat projects are used; below about five people the collaboration premium is hard to justify. Against NotebookLM (free) and a $20/mo ChatGPT Plus seat, Liminary is the more expensive choice — you are
In short
Liminary — Liminary is a browser-extension research memory layer that captures what you read and fact-checks deliverables inside Notion and Gmail. Best for Management consultants producing client deliverables across multiple sources, Strategy and freelance consultants advising several clients at once, Market researchers tracking industry trends, whitepapers, and reports. Free to start; paid plans from $29/mo.
What's new in Liminary
Checked 5 days agoAcross the latest 5 updates: 2 feature updates and 3 news mentions.
Benchmarking Liminary's AI memory layer: LongMemEval, Memora, and BEAM
Liminary publishes third-party benchmark results showing improved retrieval accuracy measured against LongMemEval, Memora, and BEAM, backing its claim of 3.4X better retrieval than ChatGPT.
On AI Benchmarking
Liminary explains its methodology and philosophy for benchmarking AI memory systems, the context behind its published retrieval-accuracy numbers.
Free Liminary for anyone in a job search
Liminary opens free access to its knowledge tool for job seekers, letting them organize research on target companies and roles at no cost.
Stop tab-hopping to write the email - Gmail drafting
Liminary can now draft and edit emails in Gmail from chat, pulling from your saved knowledge and writing in your voice. It will not hit send.
Fact Check and Gap Detection, now in Notion
Fact-check a Notion draft against your own saved sources and fix in one click, or run gap detection to surface missing content before a reviewer finds it.
What people actually say about Liminary — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
46 mentions across 4 sources (Hacker News, YouTube, Product Hunt, Lemmy) · researched Aug 11, 2026.
Average across the 4 sources that answered — each source counts once, not each post.
- +Source-grounded recall with every insight linking back to original
- +Proactive surfacing of saved research without manual search
- +One-click capture across web, PDFs, AI chats, YouTube, and email
- +Writing assistant with fact-checking and gap detection in Notion
- +Live meeting assistant fetches relevant context during calls
- −Independent, third-party reviews are rare, making trust hard
- −Free tier's 200 documents is restrictive for heavy research
- −Unclear how conflicting or outdated sources are handled
- −Pro plan at $29/mo may be pricey for individual experimenters
- −No evidence on performance at scale with thousands of documents
- • No explicit hidden costs mentioned, but meeting assistant and advanced features are gated behind Pro
- • Collaborative features may require Team plan at higher per-seat price
Viability Score
How well maintained and how widely used is Liminary? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- One-click capture of web pages, PDFs, AI chats, and videos while browsing
- Annotate sources with your own notes and context as you read
- Proactive recall surfaces relevant past research before you search
- Fact-check a Notion draft against your saved sources and fix in one click
- Gap detection surfaces missing content in draft deliverables
- Gmail drafting and editing from chat in your own voice, no auto-send
- Magnet live meeting assistant recalls knowledge during calls and saves transcripts
- Magnet saves meeting summary and notes for later recall
- Watchlist agent monitors clients and projects for new information
- Every answer labels whether it came from your data or the external web
- Source provenance: each insight links back to the original source
- Model selection on the Pro tier
- MCP Connectors and Agents, including Claude MCP server setup
- Shared team knowledge base with up to 50 people per project
- Data encrypted, siloed per workspace, never used to train AI models
About Liminary
Liminary acts as persistent memory for people whose paycheck depends on being right in front of a client. You save articles, PDFs, AI chats, and videos with one click while you browse, annotate them with your own notes as you read, and then Liminary surfaces relevant past work before you think to search for it. Set a watchlist on a client or project and it keeps monitoring for new information. Every answer tells you whether it drew on your own saved data or the external web, and every source is one click away. The workflows run from capture to deliverable: fact-check in Notion landed in May 2026 so you compare a draft against your own sources rather than the open internet and fix in one click, and Gmail drafting shipped in June 2026 so you can write or tighten an email from chat in your own voice, pulling from saved knowledge, without hitting send. Magnet, the live meeting assistant, recalls your knowledge during calls and saves the transcript, summary, and notes. Liminary's team comes from Dropbox security, ML, and search engineering; your data is encrypted, siloed per workspace, and never used to train AI models. It targets management consultants, strategy and freelance consultants, market researchers, think tank and policy analysts, executives, and VCs who bill for recommendations, not casual note-takers.
Behind the Verdict
The category Liminary competes in — AI memory, second brain, knowledge layer — is crowded and mostly dishonest about what it does. Most tools are a search box over a folder of PDFs. Liminary's actual claim is narrower and more defensible: it is the retrieval layer that sits under the chatbot you already use, and it shows its work. That distinction matters for anyone whose output gets scrutinized. When you ask a question, Liminary tells you whether the answer came from your saved sources or from the open web, and links back to the original. For a consultant putting a number in a client deck, that is the whole ballgame. What it does well: capture is frictionless (web pages, PDFs, AI chats, videos, one click, annotate as you go), the watchlist agent keeps monitoring a client or project so you are not manually re-checking, and the newer Notion and Gmail moves put the memory where the work actually happens. Fact-checking a Notion draft against your own saved sources rather than the internet is a genuinely different function from what ChatGPT does, and Gmail drafting that pulls from your saved knowledge and does not press send is the right amount of caution for a professional context. Magnet, the meeting assistant, is the natural extension — recalling your research live during a call and saving the transcript and notes afterward. The honest caveats. It is a browser extension, which means no offline mode, no native mobile app, and no on-prem deployment — a hard stop for regulated teams with data-sovereignty requirements. The free Explorer tier is a real trial rather than a working setup: 200 saved documents, 10 local uploads per day, 3 composer uses, 30 days of chat history, 10 Magnet meetings. Casual note-takers who enjoy organizing in Obsidian or Notion by hand will find Liminary redundant, and people who really just want a chatbot should use a chatbot. The 3.4X retrieval-accuracy figure against ChatGPT is the company's own number; it also published benchmark work against LongMemEval, Memora, and BEAM in August 2026, which is the closest thing to independent verification available, and worth reading before you take the multiplier at face value. Team at $50/user/mo only becomes worth it when the shared knowledge base and 50-seat projects are real, not aspirational. Where it fits: independent consultants juggling multiple clients, analysts drowning in reading, and anyone who has ever rewritten a section from memory because they could not find the source. Where it does not: mobile-first workers, offline environments, and anyone with no recurring research burden to compound.
Researching Liminary? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Liminary actually fits — and what changes day-one when you adopt it.
You spend the first morning of a new engagement saving analyst reports, PDFs, and prior AI chats with one click, annotating each as you go, then set a watchlist on the client.
Outcome: When you start drafting the kickoff deck, Liminary surfaces the competitive analysis you did six months ago for a different client, and every claim in the draft traces back to a source you can hand over.
You set a watchlist on the competitor and keep working; Liminary monitors for new articles and reports without you re-checking.
Outcome: New material is waiting for you rather than found late, and when you write the update you can fact-check the draft against your saved sources in Notion and fix errors in one click.
You move the team to the $50/user/mo Team plan, connect Google Drive and Granola, and let up to 50 people per project capture and annotate into one knowledge base.
Outcome: Every teammate's captures, annotations, and meeting notes are searchable across the workspace, so a new hire inherits the team's research instead of rebuilding it.
Use Cases
- A consultant researching a new client briefing saves web pages and PDFs, then drafts the document with automatic citations.
- A market researcher sets a watchlist on a competitor and gets alerted when new articles are published.
- During a client meeting, Magnet surfaces relevant research from past projects without manual prompting.
- An analyst fact-checks a draft report against saved sources directly in Notion and fixes errors in one click.
- A freelance consultant juggling clients recalls insights from one engagement while working on another.
- A team maintains a shared knowledge base where every member's captures and annotations are accessible across the workspace.
- A job seeker uses free access to organize research on target companies and roles during a search.
- An executive pulls prior research into a live call instead of asking an analyst to re-send it.
Models Under the Hood
as of 2026-09-23
Limitations
- Liminary is a Chrome browser extension, so there is no offline mode, no native mobile app, and no on-premise deployment — a blocker for regulated teams with data-sovereignty requirements.
- The free Explorer tier is a trial rather than a working setup: 200 saved documents, 10 local uploads per day, 3 composer uses, basic recall, 30 days of chat history, and up to 10 Magnet meetings.
- Pro at $29/mo raises that to 10,000 saved sources, 100 uploads per day, advanced recall with fact-check and gap detection, unlimited Magnet meetings, 24 months of chat history, and collaboration with up to 10 people per project.
- Team at $50/user/mo adds the shared knowledge base, Drive and Granola workflow integrations, advanced AI models, unlimited chat history, and up to 50 people per project.
- The 3.4X retrieval-accuracy claim versus ChatGPT is the vendor's own figure; it does publish benchmark methodology against LongMemEval, Memora, and BEAM (August 2026), which is the best available independent check.
as of 2026-10-02
Verification history
We have re-verified Liminary 7 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 7 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Liminary tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Explorer
$0/mo
Ideal for
First-time evaluator who wants to test the capture-to-recall loop on real work before committing, e.g. one active client project.
What this tier adds
Free entry point: 200 saved documents, 10 local uploads per day, 3 composer uses, basic recall, 30 days chat history, 10 Magnet meetings.
Pro
$29/mo
Ideal for
Solo consultant, analyst, or VC associate billing hourly who needs recall, fact-check, and unlimited meeting capture on live work.
What this tier adds
Adds 10,000 saved sources, 100 uploads per day, advanced recall with fact-check and gap detection, unlimited Magnet meetings, 24 months chat history, model selection, and MCP connectors.
Team
$50/user/mo
Ideal for
Small consulting or research firm of roughly five-plus people that wants one shared knowledge base rather than isolated personal memory.
What this tier adds
Adds unlimited sources and uploads, a shared team knowledge base, up to 50 people per project, Drive and Granola workflow integrations, advanced AI models, and a dedicated support channel.
Where the pricing makes sense
The company stage and team size where Liminary's pricing actually pencils out — and where peers do it cheaper.
Explorer ($0) is a real test of the capture-and-recall loop but not a working setup. Pro at $29/mo is priced for a solo consultant or analyst who bills hourly and gets the money back the first time it saves a re-research hour. Team at $50/user/mo only pays off when the shared knowledge base and 50-seat projects are used; below about five people the collaboration premium is hard to justify. Against NotebookLM (free) and a $20/mo ChatGPT Plus seat, Liminary is the more expensive choice — you are
Setup time & first value
How long it actually takes to get something useful out of Liminary — broken out by persona, not the marketing-page minute.
Solo consultant: install the Chrome extension, sign in, and start capturing in a few minutes; first real value (a surfaced past source) typically lands in the first working session. Team rollout: budget a short admin pass to connect Drive, Granola, and Slack, invite seats, and seed the shared knowledge base — expect a few days before cross-member recall is useful, plus custom onboarding on the
Switching to or from Liminary
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From NotebookLM or ChatGPT history: re-save the underlying articles, PDFs, and reports with one-click capture, then ask across everything instead of one project at a time.
- →From Obsidian or Notion notes: keep your existing vault, capture your sources into Liminary as you re-read them, and use Liminary for recall while keeping the notes where they live.
- →From bookmark managers and read-later apps: save links directly into Liminary with annotations so capture and recall sit in the same place.
- →From browser history and open tabs: capture the pages worth keeping as you browse, so future recall is scoped to real sources rather than everything you clicked.
- ↗To NotebookLM: export the underlying source documents and rebuild a notebook per project, accepting that you lose cross-client recall and provenance labeling.
- ↗To a general chatbot: copy your saved sources out manually and rely on the model's own knowledge — you lose the personal memory layer and the your-data-vs-web distinction.
- ↗To Notion or Obsidian alone: paste captured sources and notes into your vault, trading proactive recall for full manual control of structure.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Liminary”, and we withheld 6: 6 could not be judged, because “Liminary” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Liminary.
Official links
Tools that pair well with Liminary
Common stack mates teams adopt alongside Liminary, with the specific reason each pairing earns its keep.
Mem
Mem is an AI knowledge base that captures your notes, meetings, and research and keeps a living picture of your work current in the background.
remio
Local-first AI agent that turns your files, meetings, and research into finished decks, spreadsheets, and reports.
Paxton AI
Paxton AI is a legal AI assistant for litigation teams — tied to a matter record, research, and drafting across all 50 states.
Featured Head-to-Head Comparisons
Liminary vs Praktika
If you're a consultant or researcher drowning in sources, Liminary's auto-capture, fact-checking, and meeting sidekick make it the clear choice. If you're an intermediate language learner wanting immersive speaking practice, Praktika's AI tutors and real-time corrections are purpose-built. These tools address completely different needs—choose based on whether your bottleneck is information retention or language fluency.
Liminary vs Surge Ai
Liminary is the right choice for individual professionals who need to manage and recall research without manual effort, especially if you use Chrome and want integration with Gmail, Notion, and meetings. Surge AI is built for enterprise AI teams that require expert human feedback to train and evaluate frontier models; it is not a tool for personal productivity. Choose Liminary for knowledge work, Surge AI for AI alignment and RLHF.
Alternatives to Liminary
View all
Frequently Asked Questions
Used Liminary? Help shape our editorial sentiment research.