Every AI you use.
One memory that compounds.
AI tools come and go. What you've taught them stays yours. Every conversation you save makes the next one smarter, in every tool you use.
Graph
Full Graph • 2,042 nodes, 3,080 edges
Three memories carry the change: seat pricing held, the June churn interviews stress-tested it, and an annual plan was added after enterprise asks.
Follow a path
Clarify meanings · highlights 12 graph items
Timeline
Recent captures, discoveries, and working memory
What's on your mind?
174 memories may be worth maintaining
Mem checked 500 memories and found 174 worth reviewing. 30 are ready now; 144 can wait. Nothing has changed yet.
Current review
30
Ready to move aside
0
Later
144
Examples from this review
Pricing decision enriched with churn interview evidence
Background intelligence folded the June churn interviews into the pricing memory. Nothing was lost; v2 is kept.
confidence 0.85Your May pricing conclusion supersedes April's
Kept newer as latestMemories
Find specific memories, entities, and threads.
Pricing: seats, not usage
Interviews showed teams budget per person; usage-based bills spiked anxiety in every trial. Charge per seat, not per call.
Trial length: 14 days beat 30
Thirty days gave people time to forget why they signed up. Fourteen days kept the decision close to the aha moment.
Enterprise asked for annual invoicing
Every enterprise call this month asked for an annual invoice option instead of monthly card billing.
Launch retro: waitlist emails converted 3x social
The waitlist converted at 9% versus under 3% from social traffic. Build the list first next time.
Churn interview: onboarding, not price
Three cancellations this week all cited a confusing first session, not the price. Onboarding is the leak.
Type
Source
Library
Your wiki, derived live from every source you keep.
+12 more topics
Pricing decisions
Graph
InvestigateRecent on this topic
Seats over usage, annual added for enterprise
May 9Annual invoicing folds enterprise asks into the pricing decision without reopening the seat-vs-usage question.
Launch retro: waitlist emails converted 3x social
May 2The waitlist converted at 9%; social traffic barely cleared 3%. The ledger caught it in review.
Churn interview: onboarding, not price
May 9Cancellations trace to a confusing first session, not the sticker price.
Graph Intelligence walks your graph and shows every step
Save it once, from any tool.
Finish a session in Claude Code, Cursor, or a browser chat, and save it. Mem keeps the whole thread, word for word, and distills the decisions inside it into memories you can find later. The conversation stays. The knowledge compounds.
Thread · kept whole
So which pricing model do we go with?
Claude Code · 14 messages · Mar 3
> So which pricing model do we go with?
● Seats, not usage. Every interview showed teams budget per person.
Memories · distilled
Pricing: seats over usage
Enterprise wants annual invoicing
Your memory grows on its own.
Background intelligence links related memories, folds new evidence into old conclusions, and retires the ones newer evidence replaces. Every memory keeps its version history, so you can always see what changed, and why.
Tue, you saved a churn callMem noticed it touches your pricing decisionThe decision updated itself. History kept.
Background Intelligence
June churn interviews folded into the pricing decision
This was a routine background step. It is kept for audit, but there is nothing you need to do.
Version history (3)
Seats + annual for enterprise
Seats, teams budget per person
Usage-based, cheaper to start
Find it again.
One search across every memory, thread, and document you keep. Your agents get the same reach over MCP, scoped to what the task needs.
YouEvery source, one search
Pricing: seats, not usage
Enterprise asked for annual invoicing
Your agentsPlugin, CLI, skills, or MCP: one memory
Seats, decided Mar 3: teams budget per person, and the June churn interviews confirmed price wasn't the real objection.
Pricing: seats, not usageIt learns how you work.
Solve the same kind of problem a few times, and Mem drafts a skill from the evidence across your threads, proves it against real checks, then keeps improving it. A new version ships only if it beats the current one on checks it has never seen. The more you use it, the sharper it gets.
From your work
user-interview-debrief
1. Capture their exact words for the pain
2. Separate what they said from what they meant
3. End with one decision or one open question
Make it sharper
Used 14 times · every run reports back
Ask it something hard.
Graph Intelligence actually walks your graph: a node, its community, then the global picture, and it shows every step it takes. Ask a question, or go further: co-write the article, run the research, build on everything you already know.
The shortest path crosses two communities: trial anxiety reaches the pricing decision only through the June churn interviews.
Usage pricing tested well with power users but scared off the median trial signup, and the June churn interviews confirmed price wasn't the real objection.
One answer, receipts across three layers: memory, conversation, document.
Also surfaced: 2 cross-community links you never made by hand.
Local and Cloud, one Mem.
Mem runs on your own machine first: everything local, fully offline if you want. Nowledge Cloud is the same product running on servers we operate, so memory and background intelligence stay on while your laptop sleeps. Moving in is one click, moving out just as easy: where it runs is your choice. That it stays yours is not.
Pricing: seats, not usage
Charge per seat, not per API call. Interviews showed teams budget per person; usage-based bills spiked anxiety in every trial.
Offline right now. Everything still works.
Local
- Your memory never leaves your machine
- Works fully offline
- Install once, yours to keep
Morning summary
Two pricing memories linked overnight
Background intelligence kept working while the laptop slept.
3:12 AM. Still working.
Nowledge CloudComing soon
- The whole Mem in a browser tab
- On 24/7, working while you sleep
- Shared workspaces for teams
Get Nowledge Mem.
Free to download, local first. Bring your own AI keys, or add Plus: managed AI in one subscription. A cloud workspace that never sleeps is on the way.
Already have an AI tool? One URL connects it: mem.nowledge.co/SKILL.md