For people who think with AI

The permanent memory layerfor AI conversations.

Save responses, highlights, and full conversationsfrom ChatGPT, Claude, and Gemini. Permanently.

Free to install · Works on ChatGPT, Claude & Gemini · Local-first, private by design

Archive
0 saved
0 entries · 0 topics · 0 AI modelsLocal-first · Private

You’ve already lost
thousands of insights.

Not because you forgot. The toolsweren’t built to remember.

You researched it in Claude. Worked through it in ChatGPT. Found the exact framing in a conversation from three weeks ago.

It’s gone now. You’ll do it again.

AI conversations are where serious thinking happens today. Research, decisions, drafts, breakthroughs. But they were designed to be temporary. Every session ends. Every insight fades.

The cost isn’t obvious. It accumulates quietly.

What Savelore is

Not a tool.
Knowledge infrastructure.

Most AI tools are stateless. Every session starts from zero. Savelore introduces continuity. The missing layer between your thinking and the tools you think with.

Permanent

Responses, highlights, and full conversations. Saved forever.

Searchable

Full-text search across every entry you've ever saved

Editable

Trim saved content down to exactly what you need

Private

Local-first by default. Your archive never leaves your device unless you choose

This is what continuity feels like.

The system

The system that makes your thinking permanent.

Every layer works together. Nothing falls through.

01 · Capture

Save exactly what matters. Nothing more.

Save a single response, a highlighted excerpt, or a full conversation from ChatGPT, Claude, or Gemini. Even temporary and incognito chats. Each entry lands in your archive permanently, with a direct link back to the exact message.

Responses, highlights & full convos · incognito too
Claude · Architecture Discussion
U

What's the best approach for handling rate limits in a distributed API?

AI

02 · Organize

Your archive reflects how you think.

Assign folders, tags, and a personal description your way. Filter by model, date, favourites, or pins. Your structure, not ours.

Manual organization · multi-model
Archive
0 saved
0 entries · 0 topics · 0 AI modelsLocal-first · Private

03 · Retrieve

Find it in seconds, not hours.

Full-text search across every entry, description, and tag you've ever saved. The more you save, the more powerful your search becomes.

Full-text search · entries, descriptions & tags
Search your archive
47 entries
Search your memory…

04 · Synthesize

Stitch research from any AI into one entry.

Collection Mode lets you highlight prompts and responses across different conversations, even across Claude, ChatGPT, and Gemini, and stitch them into a single cited entry. Every citation links back to the exact message it came from.

Cross-platform · cited sources · jump to message
Collection Mode
Stitching…Entry ready
2 platforms

Collecting highlights

Claude[1]

The token bucket algorithm is preferable here because it allows short bursts while maintaining a long-term average rate. Unlike leaky bucket, which enforces strict uniformity.

ChatGPT[2]

Redis is the standard backing store for distributed rate limiting. For fallback: keep an in-memory counter per instance. Accept slight over-serving during Redis unavailability rather than blocking all traffic.

Claude[3]

Implement rate limits at the API gateway layer, not in individual services. The gateway sees all traffic. Services only see what gets through.

Saved to archive

3 sources · 2 platforms

Rate Limiting Strategy: Distributed Systems

Token bucket algorithm preferred for burst tolerance while maintaining average rate.

Redis as backing store; in-memory fallback during outages. Accept slight over-serving over blocking.

Enforce at the API gateway layer, not per-service.

Sources

[1]
Claudejump to exact message
[2]
ChatGPTjump to exact message
[3]
Claudejump to exact message

05 · Compound

The more you save, the more you can find.

Every saved entry makes your archive more complete. Six months of saving means six months of thinking instantly searchable. Whenever you need it.

Your archive · more complete over time
Your archive
6 months of saving
6 months agoviewing

Initial database schema

Decided on PostgreSQL with JSONB columns for flexible metadata storage...

Architecture
3 entries tagged Architecture
3 months agoviewing

API design principles

RESTful with resource-based URLs. Rate limiting per user token, not IP...

Development
3 entries tagged Development
3 weeks agoviewing

Scaling decisions

Revisiting the database schema from January. JSONB is proving flexible enough...

Architecture
3 entries tagged Architecture
Just nowviewing

Performance audit

Query patterns that reference the original schema and the API design...

Development
3 entries tagged Development
4 entries · searchable instantly
Architecture · Development

Your data. Your archive.

Your thinking should belong to you.
Fully. Permanently.

We built the architecture around that belief. Not as a compliance checkbox, but as a decision made before the first line of code.

Savelore is local-first. Your archive lives on your device by default. We offer optional sync for access across devices. Your choice, your control.

We don’t sell your data. We don’t train on your conversations. We don’t share it with third parties.

Your archive is private because an archive that isn’t private isn’t really yours.

Private by architecture. No exceptions.

Architecture

Your deviceArchive stored locallyoptionalOptional syncYour choiceNowhere else.No third-party accessno advertising · no data brokering

Pricing

Start free. Upgrade when you’re ready.

Free gives you the full archive experience, locally. Pro removes the limits, adds sync, and connects to the tools you already use.

Free

$0forever

No credit card required.

  • Save responses, highlights & full conversations
  • Folders, tags, filters & favourites
  • Jump to exact message in source conversation
  • Full-text search across all entries
  • Local-first. Private by default.
  • Up to 10 saved entries
  • 2 full conversation saves
  • 2 Collection Mode uses
Add to Chrome. Free.

Pro

7-day free trial
$9.99/ month

No charge for 7 days. Cancel any time.

  • Unlimited entries
  • Unlimited conversation saves
  • Unlimited Collection Mode
  • Cloud sync across all your devices
  • Web archive at web.savelore.app
  • Notion export & auto sync
  • Export as PDF, HTML, TXT or JSON
  • JSON import
  • Priority support
  • Early access to new features

To upgrade:

  1. 1Install the extension
  2. 2Sign in or create an account
  3. 3Navigate to Settings. Tap Subscription.
  4. 4Start your free trial.

A verified account is required to subscribe to Pro. Your archive stays on your device by default. Sync is a Pro feature.

Where this goes

The archive is only the beginning.

Today Savelore captures, structures, retrieves, and compounds your thinking across any AI tool.

What comes next: a memory layer sophisticated enough to surface what you knew three months ago when it’s relevant to what you’re working on today. Semantic retrieval. Intelligent resurfacing. Connections that form across time without you having to make them.

Not AI doing your thinking.
AI helping you hold onto it.

Your archive today. Your memory layer tomorrow.

No hype. No dates we can’t keep. Just the work.

Begin your archive

Every conversation you have

today could still be

with you in five years.

Or it could be gone by next week.

Savelore is the difference. Your archive starts small and compounds forever. Every conversation a permanent layer in a system that grows with you.

Free to install · No credit card required · Local-first, private by design

Local-first. Private by design.

Built for the long term.

Your archive belongs to you.