Lore.AI — Shared Context for AI Agents

loading…<br>Lore. The memory that compounds.

Stop re-explaining

your project to<br>your AI.<br>Your team's memory, in every session. Lore gives AI agents persistent shared context — capturing<br>decisions, file paths, and patterns across sessions lasting days and hundreds of turns.<br>No context files to maintain. No workflow changes.<br>{this.querySelector('.install-copied').classList.add('show');setTimeout(()=>this.querySelector('.install-copied').classList.remove('show'),2000)})"> $ curl -fsSL https://withlore.ai/install | bash Copied!<br>or npx @loreai/gateway

Join the Waitlist<br>View Repository Read Docs

Lore.AI Gradient Context<br>Lore Distillation<br>Any Provider*<br>Recall Tool<br>.lore.md Sync<br>On-Device Vector Search<br>Import History<br>Cost-Aware Caching<br>Sessions Lasting Days

+67%<br>vs Compaction at 2.3M Tokens

2.6x<br>Total Recall vs Compaction

2.3M+<br>Token Sessions Tested

2.3M tokens, 5 days, 2.6x total recall ◆ Compaction: 2.4/5. Lore: 4.0/5. 68 min/day re-explaining ◆ Lore remembers for you Your tools change. Your memory doesn't. ◆ Lore is your constant Total amnesia on new sessions ◆ Lore persists across sessions 49 manual learnings ◆ Lore curates automatically 5 feedback loops ◆ Your agent improves every session 2.3M tokens, 5 days, 2.6x total recall ◆ Compaction: 2.4/5. Lore: 4.0/5. 68 min/day re-explaining ◆ Lore remembers for you Your tools change. Your memory doesn't. ◆ Lore is your constant Total amnesia on new sessions ◆ Lore persists across sessions 49 manual learnings ◆ Lore curates automatically 5 feedback loops ◆ Your agent improves every session

The Problem<br>Context loss is invisible.

There's no error message when your AI forgets.<br>Just worse answers, undone decisions, and<br>hours spent re-explaining.

01<br>Compaction destroys details<br>When the context window fills up, your AI tool<br>compacts the conversation.<br>In a real 5-day coding session, compaction reduces 2.3 million tokens to an 11K summary — a 200x<br>compression that loses which issues were picked, what alternatives were rejected, and why.<br>It scores 2.4/5 on recall. Lore scores 4.0/5.

02<br>Starting fresh is starting from zero<br>Most developers see "Compacting conversation" and start a new session. That<br>trades compaction for total amnesia.<br>The new session produces output that looks fine — but it's working<br>from incomplete information, and you can't tell.

03<br>Manual context files don't scale<br>The alternative is maintaining context files, key technical learnings, and decision<br>rationales — by hand. It works, but it's a second full-time job. One team<br>tracked 49 technical<br>learnings manually. Every decision needs the "why" or the AI will refactor it away.

The Solution<br>How Lore replaces all of that

01<br>Intercept<br>Lore sits between your AI client and the upstream API. It captures every message — no<br>client changes needed, just change the base URL. Works with Claude Code, OpenCode, Pi, Codex,<br>and any Anthropic/OpenAI-compatible tool.

02<br>Distill<br>Lore replaces compaction entirely. Instead of lossy summaries that forget your file paths<br>and decisions, it distills conversations into timestamped observation logs — the operational details<br>your AI actually needs to keep working. Your manual "Key Technical Learnings"? Lore extracts and<br>maintains them automatically.

03<br>Recall<br>Details from every session are searchable — even hundreds of turns later.<br>When the distilled context isn't enough, your agent's recall tool retrieves the exact file path,<br>error message, or decision rationale it needs. In our 2.3M-token benchmark:<br>2.6x total recall over compaction — 13 perfect scores vs 5.

Why not both?<br>Context management and memory are the same problem.

Other tools force you to solve them separately.<br>Lore treats them as one continuous pipeline.<br>See how Lore compares &rarr;

01<br>Memory alone isn't enough<br>Storing past conversations and searching them later is only half the problem.<br>If your AI still gets compacted mid-session and loses track of what it's doing right now,<br>a memory layer can't help — it doesn't know what's missing until you ask. Memory is only<br>useful if it reaches the AI at the right time.

02<br>Context management alone doesn't learn<br>Compressing conversation history keeps the current session alive, but nothing<br>is extracted from the compression. Start a new session and you're back to zero. Switch tools<br>and the knowledge stays behind. Nothing transfers to other projects, team members,<br>or even other models.

03<br>Lore connects them into one pipeline<br>In Lore, context compression is the memory pipeline. Distillation<br>feeds the gradient context manager, which feeds the knowledge curator, which feeds<br>.lore.md<br>— and with Folk Lore, your team. Every conversation makes every future session smarter,<br>across any provider, any tool, any team member. Every new session starts with the relevant<br>facts and gets a fresh injection after the first turn.<br>Read the docs &rarr;

Persistence<br>Decisions stick<br>Your AI won't refactor away deliberate decisions. Lore preserves the "why"...