ai-memory 2.0.3 — a patch release hardening the 2.0 line: data-safety, provider correctness, and reliability. No breaking changes; drop-in over 2.0.x.
Highlights
- 1.x rollback restored (#633). The pre-migration safety archive is now taken before the SQLite schema is migrated, so it is a genuine pre-2.0 recovery point a 1.x binary can reopen — the documented "reversible upgrade" works again. (Earlier 2.0.x archives captured the already-migrated DB and were 2.x-only.)
- Typed edges work on OpenAI-family providers (#630).
relationswas an open map that OpenAI strict mode silently emptied; it is now a fixed-shape object, socauses/fixes/contradictsedges emit again onopenai,openai-oauth,copilot,opencode, andopenai-compat. - bootstrap reliability. Tolerates a chunk that returns no pages (#614); retries a chunk on transient LLM errors — 5xx/429/timeout — instead of discarding the whole run (#617).
finalize-session --agentaccepts every captured agent (#623) —hermes,claude-desktop,crush,other.- Generated TypeScript integrations authenticate again under
--apply(#625). - Watcher no longer spams on orphan project dirs (#613, #616); dangling-project error clarified (#612); antigravity Windows hook quoting fixed (#611).
- Operability:
ai-memory statusand the pre-migration log now surface filesystem free space (#629); release/Docker/cargo installbuilds are--locked(#628); the auto-improve pending sidecar drops its always-pendingstatus line (#624).
⚠️ Upgrading from 1.x
The OKF v0.2 wiki-format migration is automatic and backup-gated. As of 2.0.3 the automatic safety archive is a true pre-migration recovery point (restore it and start the old 1.x binary — #633). Full steps, including the surgical git-checkpoint rollback: docs/MIGRATION-2.0.md.
Full change list: CHANGELOG.md.
Checksums (SHA256)
0f547a0057c2a15a3f57ede4087215f905765c4ebaadbc60e00852b97ef0aa34 ai-memory-hooks.tar.gz
448f6cfad616e13692598dd49709ce9dec10eed0f702818ccca4a0c54a79f8c7 ai-memory-install-hooks
c053935c88c0b3f22a28a9a138fdbdbc9eb3ca5aa8626a3128fb56611460d32e ai-memory-linux-aarch64.tar.gz
bdc2ae7623eb3bf39eb446f4c2b51aa65e60ce4496ac2c4c9618d14f59e464c9 ai-memory-linux-x86_64.tar.gz
466498c3dfb59b495d5ec232622357215a571dccbe4494cc3a03afa98251143f ai-memory-macos-aarch64.tar.gz
6a47c4078215b62fd0f190892fb07cce08b6ae88dbaed324a4310b9dffd56bd6 ai-memory-macos-x86_64.tar.gz
f5614fe3da119696ef31182032dc09cb145f71efcc8a401c50a339542a52e9e8 ai-memory-windows-x86_64.zip
6a91c44ffa2e85d3b5a6ce26a4ccff0286b91d08d276d2bb4d47c346b036ee8b ai-memory-wrapper.cmd
2d35c2f1fba28a1aee3ffeb58e3deaec772803f3dc6fd159de6ae0a47ba797ac ai-memory-wrapper.ps1
4b2e5736195f0ac4cd38adf62f25b294922ebf81f5f4802a07803e1abbf9f72d ai-memory-wrapper
Install
# Arch Linux (AUR)
yay -S ai-memory-bin # prebuilt Linux x86_64/aarch64 binary
yay -S ai-memory # builds from source
# Docker
docker pull akitaonrails/ai-memory:2.0.3
# macOS aarch64/x86_64 (no toolchain): download the matching ai-memory-macos-<arch>.tar.gz below,
# extract it, and follow the bundled docs/macos.md.
# Windows x86_64 (no toolchain): download ai-memory-windows-x86_64.zip below,
# extract it, and follow the bundled docs/windows.md.
# From source
cargo install --locked --git https://github.com/akitaonrails/ai-memory --tag v2.0.3