Capture what matters
Save meeting notes, decisions and project details in a Space you choose. Track the processing before asking about them.
Capture a notePrivate preview · Persistent AI workspace
Keep the context.
Move the work forward.
A persistent AI workspace that connects project knowledge with ongoing work. Research, create content and coordinate coding tasks with context.
Existing accounts can sign in. Public registration is currently closed.
- Project Spaces
- Source references
- API + CLI + MCP
Project AuroraExample workflow
01 · Capture a decision
We chose PostgreSQL because we need transactions and relational queries.
02 · Process the note
Check its progress in Processing.
03 · Ask later
Why did we choose PostgreSQL?
For transactions and relational queries.
Source: the project’s database decision.
Your workspace
From a saved detail to useful work.
Start with one project and one note. Add sources and connections as your workflow grows.
Ask with context
Ask questions about your stored information. Keep the source references beside the answer and continue earlier conversations.
Open AskRead the public web
Read articles and documentation directly. A small crawl can follow same-site links, without a paid crawler for ordinary HTML pages.
Read a websiteSeparate your projects
Keep work, personal notes and shared projects in their own Spaces. Choose the scope before you capture or search.
Manage SpacesSee what is happening
Check queued work, completed steps and failures. Accepted for processing is different from ready to search.
Check processingStay in your workflow
Capture, search and ask from the CLI. Connect your application through the API or an assistant through MCP.
Read the CLI guideAnimated explainer · 49 seconds
Keep the context. Keep creating.
Follow an animated founder from scattered ideas to connected knowledge. See the real web scraper and a saved project decision in PersistMemory.
Original character animation and real product recordings with sample data. Includes AI narration, original music, sound effects and captions.
Open the videoInteractive example · sample data
Decisions change. Keep their history.
Choose a sample question to explore how current information and earlier decisions fit together. This example does not access a real account.
Pick a question
Four memories match, and every one of them is real
Ranked by similarity
24 Aug0.95
Team later moved to another provider.
1:1 with John
18 Aug0.90
Bedrock implementation shipped.
Standup notes
10 Aug0.81
Decided: Amazon Bedrock.
Planning doc
2 Aug0.62
Bedrock raised as an option for model access.
#eng-infra · Slack
A plain search returns
The best-matching line.
Team later moved to another provider.
1:1 with John · 24 Aug · 0.95
End of result
True, and it does not say what was moved away from, when it was chosen, or that anything shipped on it. The three memories that explain the sentence scored lower and were left behind.
PersistMemory returns
All four, and which one is now.
RELEVANT MEMORY
CurrentTeam later moved to another provider.
1:1 with John · valid: 24 Aug — still true
NO LONGER TRUE — these WERE true and have since changed
Bedrock raised as an option for model access.
#eng-infra · Slack · valid: 2 Aug — 10 Aug
Decided: Amazon Bedrock.
Planning doc · valid: 10 Aug — 24 Aug
Bedrock implementation shipped.
Standup notes · valid: 18 Aug — 24 Aug
Kept in the order they happened, so a question about August is still answerable in September. Nothing was overwritten to get here.
And so the assistant answers
Not Bedrock any more. It was chosen on 10 August, shipped on the 18th, and replaced on the 24th.
Built for developers
The same memory.
Where you already work.
Use the website for everyday work and the CLI for your terminal. Bring memory into an application with the SDK or connect a compatible assistant with MCP.
Terminal · signed-in account required
pm spaces create "Project Aurora"
pm remember "We chose PostgreSQL" \
--space "Project Aurora"
# Check the returned job before searching
pm jobs get <job-id>
pm ask "What database did we choose?" \
--space "Project Aurora"For software teams · API integration maintenance
What changed in the API you depend on?
PersistMemory helps small software teams maintain external API integrations. It connects watched documentation with selected GitHub repositories and branches, investigates relevant changes, runs repair workflows in isolated environments, and prepares draft pull requests with test evidence for review.
Explore API change workflowsBefore you begin
A few practical details.
What is PersistMemory?
PersistMemory is a persistent AI workspace that connects project knowledge with ongoing work. Users organize documents, websites, conversations and decisions into permission-controlled Spaces. AI agents use that context to research, create learning materials and content, and coordinate coding tasks through the web, connected AI tools and messaging channels. The workspace brings together project memory, task history and supporting evidence so users can continue work across sessions without repeatedly rebuilding the context. It is software, not persistent-memory storage hardware such as PMEM or NVDIMM.
Can I sign up now?
PersistMemory is in private preview and public registration is closed. Existing users can sign in. Documentation and examples are public.
Does saving a note make it immediately searchable?
Capture starts background processing. Check Processing to see the job state; extraction and search indexing can finish separately.
Can it read every website?
The direct reader supports public HTML and text pages with page and size limits. Sites needing JavaScript, sign-in or other access may not be readable.
How can I connect an AI application?
Use the API, JavaScript SDK, CLI or MCP connection described in the documentation. Account access is required for your own memory.
Private preview