Rethread captures your preferences, decisions, and context from AI conversations — then brings it all back when you need it. Across six platforms, synced end-to-end encrypted across every device you own.
Silently observes your conversations and extracts key facts, preferences, and decisions. No manual tagging required.
Search, filter, sort and edit memories before injecting context. Live token estimate, summary vs detailed mode, and per-memory feedback (helpful / hide) make every recall sharper than the last.
Context captured in ChatGPT is instantly available in Claude, Gemini, Grok, Perplexity, and DeepSeek. Your knowledge graph follows you across six platforms.
Sync memories across every device with AES-256-GCM encryption derived from a passphrase only you know. Servers only ever see ciphertext — by design, even we can't read your data.
Browse 117 hand-vetted prompts across coding, writing, research, and productivity. Star, install, or publish your own to the community library.
Every morning at 8 AM Rethread builds a structured digest of yesterday — conversations, decisions, open questions, follow-ups, and top topics — surfaced as a banner and (optionally) a system notification.
Search across every captured message body, not just structured memories. Inject ranked snippets with conversation context alongside the rest of your recall payload.
Save immutable copies of any conversation at a point in time. Compare any two snapshots — or a snapshot vs. live — side-by-side with added / removed / modified highlights.
Multi-select memories or conversations, then bulk delete, assign buckets, add or remove tags, move folders, or export Markdown — all in a single sweep.
Alt+Shift+R recalls context on the active page. Alt+Shift+P opens the floating prompt picker. Alt+Shift+L opens the side panel — without leaving your keyboard.
Group memories and conversations into color-coded buckets, folders, and dynamic tag filters. Combine with time-range search to slice context any way you need it.
Local-first by default. Cloud sync is opt-in and zero-knowledge — keys derive on-device via PBKDF2 → HKDF → AES-256-GCM. No analytics, no telemetry, no tracking.
Use ChatGPT, Claude, Gemini, Grok, Perplexity, or DeepSeek as you normally would. Rethread works silently in the background, extracting meaningful context from your conversations.
Your preferences ("I prefer TypeScript"), decisions ("We chose PostgreSQL"), and facts ("I'm a backend engineer") are captured, tagged, and organized into buckets. Bulk-import past chat history, then turn on optional end-to-end encrypted cloud sync to keep every device in lockstep without leaking a single plaintext byte.
Hit Alt+Shift+R or click Recall to open Selective Recall — search, sort, edit, and pick exactly the memories you want injected. Wake up to a Daily Brief at 8 AM. Browse 117 community prompts. Snapshot any conversation and diff it later.
Not raw chat logs — distilled context.
Start free. Upgrade when you need more power.
Free is local-only. Pro adds optional zero-knowledge cloud sync — your encryption keys derive on-device and never leave your browser.
Rethread is built with a zero-knowledge architecture. Even when you turn on cloud sync, our servers only ever see ciphertext — your encryption keys never leave your browser.
Install Rethread and give your AI conversations lasting memory.
Why ChatGPT's built-in memory falls short, and how Rethread captures every preference, decision, and project detail across every chat.
How Rethread builds one shared memory across ChatGPT, Claude, Gemini, Grok, Perplexity, and DeepSeek — with end-to-end encrypted device sync.
A technical deep dive into Rethread's PBKDF2 + HKDF + AES-256-GCM cloud sync architecture and threat model.
From DOM-level capture to Selective Recall — the full four-stage pipeline that turns conversations into structured memories.
A complete breakdown of every failure mode of LLM memory in 2026 — and how to build a memory layer that actually works.
Four context-engineering patterns — Persistent Profile, Project Brief, Decision Log, Just-In-Time Recall — with concrete examples.
Side-by-side review of built-in ChatGPT Memory, MemGPT/Letta, Mem0, LangChain memory, and Rethread — privacy, scope, and multi-platform support.
15 answers covering encrypted cloud sync, license activation, multi-device install, cancellation, the 14-day refund policy, and more.
Or browse the full Rethread blog for more articles on AI memory and productivity.