Intelligence is rented.Memory is owned.
Geniffy is the brain behind your AI. It keeps where your work stands, so every AI you use picks up where you left off.
14 days of Pro free · No card · Export or erase any time

Each project now, what you promised and by when, and what happened: the brain on one page.
Works inside the AI you already use
- Claude
- ChatGPT
- Gemini
- Cursor
- VS Code
- LangChain
- CrewAI
- Vercel AI SDK
- n8n
- Zapier
- Make
- Pipecat
- Pydantic AI
- LiveKit
- Haystack
- Perplexity
- Mistral
01The brain
Not a pile of facts. Where you left off.
Other memory keeps loose facts and hands back the nearest ones. Geniffy reads whole conversations and work sessions, and keeps what a person would.
AskedWhere did I leave off on billing?
- Uses Postgres.
- Works on billing.
- Priya is a customer.
The brain- NowBilling: webhooks done, the emails are left.
- DecidedRenewals moved to the new provider on 7 Oct, because the old one could not retry a failed card.
- Due FridayThe demo you promised Priya.
- LessonThe last deploy failed on a missing secret. Check the task's settings first.
Where things stand. Every project's now, next and still open, and what was decided and why.
Promises. What you said you would do, and by when. Mark it done, or see it before it slips.
What happened. Whole conversations and agent sessions, tool steps included, read as stories: what was tried, what failed, what worked.
Lessons. What worked and what didn't, kept so the next session doesn't repeat a mistake.
Every night. It settles what disagrees, shows each change with its reason and an undo, and tests itself.
The briefing. All of it in one call before any AI starts. Read, not generated: instant, and free on every plan.
- Claude, ChatGPT and Cursor, with one sign-in
- Claude Code, Codex and OpenCode: whole sessions
- Your own app, through the API and SDKs
- Files, pages and your AI history
- Where things stand, promises, what happened, lessons and facts
- Every night it settles, merges and checks itself
- The briefing, before every session
- Answers with their sources, or no answer
- Search, export and erase
- Every AI reads the same brain
Every AI opens with a briefing: where things stand, what you promised and by when, and what happened, read from your own conversations.





02Prove
Answers with receipts. Not guesses.
Every answer shows the line it came from, who said it and when. When nothing supports an answer, Geniffy says so.
Ask your memoryChronology. It knows what was true, and when. A change of plan replaces the old fact, and the old one stays on record.
Restraint. No source, no answer. Ask about something nobody said and you hear that, not a plausible invention.
Met the Lumen team today. Priya runs product there, and she wants the pilot criteria before the demo.
Lumen Labs demo prep · 3 OctProvenance. Open any memory to see the exact words it came from, then correct it or forget it.
Benchmarks
Measured, not promised.
Public memory benchmarks, judged the way the field judges them.
03For developers
Give your agents a past.
The same memory, as an API. Write what each of your users says. Read what matters, with where every line came from.
1from geniffy import Geniffy
2
3client = Geniffy() # reads GENIFFY_API_KEY
4mem = client.space("customer_1042") # one memory per user
5
6source = mem.memories.add(
7 "Priya Nair signs the Lumen renewal. It moved from March to June."
8)
9mem.sources.wait(source.id) # learning takes a few seconds
10
11print(mem.context("Who signs the renewal, and when?"))
One space per user. Every call stays inside it, and a key can be limited to one person. Nothing else can read it.
Four ways to recall. briefing() to open a session where it left off, context() for your prompt, ask() for an answer in words, search() for the memories.
Erase on request. Export everything held for a user, or forget it, in one call.
Correct and forgetWorks with your stack.
41 integrations, each one tested, each with its own page.
- AnthropicModel
- OpenAIModel
- GeminiModel
- LangChainFramework
- LangGraphFramework
- LLlamaIndexFramework
- OpenAI Agents SDKAgents
- Vercel AI SDKFramework
- CrewAIAgents
- Pydantic AIAgents
- Google ADKAgents
- MMastraAgents
- PipecatVoice
- LiveKitVoice
- HaystackFramework
- n8nAutomation
- ZapierAutomation
- MakeAutomation
- GmailData source
- Google DriveData source
- NotionData source
- SlackChannel
- WhatsAppChannel
- +18More integrationsTwilio, Zendesk, Linear and more
What you can build.
- 01An agent that knows the customer
Their plan, their last three tickets and what they were promised, before the first reply.
- 02No more "can you repeat that"
What a customer told the bot on WhatsApp is known on the next call.
- 03Hand-offs with history
A human picks up with every fact and the line it came from.
04Security and privacy
Private by design. Yours to take back.
Your memory is never sold, never shown to advertisers and never used to train a model. Take it with you or erase it, whenever you like.
Isolated per person. Each person's memory lives in its own space. A key can be limited to one person, and nothing else can read it.
Encrypted everywhere. In transit and at rest, with daily backups kept for 35 days in the same region.
Export or erase. Download everything as your own copy, or forget it: gone at once, and from backups within 35 days.
Sign-in you control. Every AI app connects through OAuth with its own permission, read-only if you like, revoked in one click.
Certifications
What is in place today, and what is on its way. A status changes only with the auditor's report in hand.
Enterprise
Your own contract and a security review today. Single sign-on and audit logs next.
Built with the companies that build AI.
- Anthropic
- OpenAI
- Google Gemini
- Amazon Web Services
- Google Cloud
What's new.
Every update- Your users' data sources, kept in memoryTested syncs for Gmail, Google Calendar, Drive, Notion, OneDrive, GitHub, Linear, Confluence, Amazon S3, Zendesk and any website.
- Claude's memory tool, kept in GeniffyThe notes Claude's memory tool writes are kept in each user's memory, exactly as written, and recalled like anything else.
- A copy for each userexport() returns every memory held for one of your users, with the sentence it came from: the answer to "what do you hold about me?"
Start free. Upgrade when it matters.
Compare every plan- Free₹0foreverTry the brain on your own work
- Learns from 10 hours of AI work a month
- Two AI apps save, every app reads
- Unlimited briefings and search
- ProMost popular₹1,499a monthOne brain for every AI you use
- Learns from 80 hours of AI work a month, 8x Free
- Every AI app saves and reads
- Answers with sources, in every app
- Max₹7,999a monthFor agents that work all day
- Learns from 400 hours of AI work a month, 5x Pro
- Everything in Pro
- Built for agents running all day

Begin with everything you know.
Start with 14 days of Pro, free. The brain behind every AI you use, with a receipt for every answer.
- Measured on Geniffy's memory engine test harness in September 2026: LoCoMo, MSC and ABCD conversations and synthetic cases, 190 to 1,373 questions per figure, judged with the LongMemEval judge prompt.

