🦛 Hippo

Stop re-teaching your agent.

Make your agent's memory work like a brain. Hippo is long-term memory for coding agents. It's a critical layer for your AI harness that connects across your different tools (Cursor, Claude Code, Codex). It keeps your proprietary data completely local, and it actually learns over time. By strengthening memories each time they are recalled, Hippo preserves what works, lets mistakes decay, and continuously compounds your agents' intelligence.

Hippo keeps your coding agents' memories in a SQLite store on your machine, with markdown mirrors you can read and commit. Search is BM25 out of the box, with no model and no network call; embeddings are an optional install. hippo init installs hooks for Claude Code and OpenCode, adds 2 hooks to Codex's hooks.json when Codex is installed (Codex runs them once you trust them in /hooks), and adds instructions to an existing AGENTS.md for Codex, Cursor, OpenClaw and Pi. Any MCP client can connect. Mark a memory wrong and it ranks lower; run hippo supersede and the old fact leaves recall. Zero runtime deps.

Install it, then run hippo init inside one project. Init creates the project's .hippo/ store and adds a block to the CLAUDE.md or AGENTS.md already there. On your machine it adds hooks for the agents it finds, such as Claude Code's in ~/.claude/settings.json, and a daily 6:15am run. What hippo init changes lists all of it and the flags that skip each part.

npm install -g hippo-memory && hippo init

Setting up every git repo under a folder in one go is a second step. The Quick start says what it changes, then gives the command.

Package installation alone does not enable automatic preservation on every agent. Complete the documented setup and required host trust; capture and compaction coverage depend on the integration.

Having an AI agent install it? Point it at llms-install.md: it installs, wires hippo into the agents it finds, and verifies with hippo doctor.

Works with:    Claude Code, Codex, Cursor, OpenClaw, OpenCode, Pi, any MCP client
Imports from:  ChatGPT, Claude (CLAUDE.md), Cursor (.cursorrules), Slack, markdown
Storage:       SQLite backbone with markdown mirrors. Git-trackable, human-readable.
Dependencies:  Zero runtime deps. Node.js 22.16+. Optional embeddings: bring-your-own local Transformers.js (`npm i @huggingface/transformers`, or legacy `@xenova/transformers`) or an opt-in API embedder (OpenAI/Voyage/Cohere). Nothing is auto-installed.

Contents: Why · Receipts · Quick start · Agent setup · MCP server · How it works · Features · CLI · Comparison · Benchmarks · FAQ · Contributing

---

Why this exists

Most "AI memory" systems save everything and search later. That's storage with search on top. A note that turned out wrong ranks the same as one that held up, and an old fact sits beside the one that replaced it.

Hippo learns from outcomes. When a recalled memory turns out wrong, mark it bad and it drops out of the top results. Memories you keep using get stronger. Those two are the parts we measured helping retrieval, on a synthetic test (mechanism audit, round 2). When a fact changes, run hippo supersede and the old version leaves recall; we have not measured whether that helps. The design borrows from the hippocampus (decay, three layers, sleep consolidation), but that is inspiration. In our tests, decay tied with decay switched off and sleep lowered recall (Benchmarks).

It also fixes the portability problem. Your ChatGPT memories don't travel to Claude. Your .cursorrules don't travel to Codex. Hippo is one store behind every agent. CLAUDE.md, Cursor rules, ChatGPT exports, Slack history, all in one SQLite store, all queryable from any tool that speaks MCP or HTTP.

---

Receipts

Numbers, not adjectives. Every claim links to the benchmark or the test that proves it. Every measurement we have ever published is indexed in docs/evals/, pre-registrations kept next to their results, including the runs that failed and the one claim we retracted.

• Sequential Learning Benchmark. benchmarks/sequential-learning/. 50 tasks, 10 buried traps. Measures whether agents learn from past mistakes, not just retrieve text. v0.11.0 informal magnitude RETRACTED v1.7.9; mechanism remains shipped. See CHANGELOG.md v1.7.9 entry.
• LongMemEval oracle split, all 500 questions in one pooled store, BM25 only, no embeddings, v0.11: R@5 = 74.0% (benchmarks/README.md, scripts in benchmarks/longmemeval/). A different setup from the per-haystack results under Benchmarks, so the two are not a before and after.
• On a private 300-query developer store, R@1 0.41 to 0.62 with hippo recall "<query>" --reranker jev, the free local cross-encoder against Jev (full eval). The opt-in TypeSafe Jev reranker, off by default, about 0.0004 USD a recall. 2000-draw paired bootstrap; the margin held in 20 of 20 seeds and a permutation null reached it in 0 of 200 runs. Ranking only: three graded tests on one 150-question LongMemEval set did not show a better answer rate than the free local cross-encoder, and that negative result is in the same doc. What it buys today is a shorter context: on that set, 2 memories ranked by Jev answered as well as 5 ranked by the cross-encoder.
• Staged Slack corpus, 10 incident scenarios: recall beat transcript replay in 10 of 10 (benchmarks/e1.3/). The answers sit mid-ch