Choose theme
AI Resources
ai-memory
ai-memory is a self-hosted memory service for carrying project knowledge between coding-agent sessions. It connects supported agents to a shared store of observations, readable notes and handoffs.
Its source of truth is a Git-backed Markdown wiki. A separate SQLite index supports retrieval, while lifecycle hooks can capture work and supply context to a later session. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
Project memory shared across agents
A Rust binary runs the service around existing coding clients. MCP connections and client-specific lifecycle integrations provide access to stored knowledge; the model and coding agent remain separate parts of the setup.
Why it stands out
Readable files remain the authority
The Markdown pages are the primary record, and the database is a derived search index. A reader can inspect or edit the wiki as ordinary files, with Git keeping versions of the stored pages.
Availability
Source and platform-specific releases
The project publishes source, release binaries and installation guides. It declares an MIT license; see the main repository for current terms. Integration depth varies by client, so the support matrix is part of choosing a setup.
Why it matters
What makes it useful
When returning to unfinished coding work, the missing piece may be the reasoning behind a choice, a failed approach or an open question. The documented handoff flow carries that context from an ending session to the next supported agent in the same directory. Stored observations and wiki search provide another place to look when a short handoff leaves out a detail.
What to know
Where it fits
The service can run on the same machine as the coding tools or on a server you operate. The setup command can create a data directory, arrange a background service and wire MCP, hooks and skills for selected detected agents. The installation guide documents a plan-only preview. These are persistent changes to the agent environment, so the selected integrations matter.
Notable points
What stands out
A handoff transfers work to a later session; it is not a live message bus between agents still running. By default, session start can claim a pending handoff automatically. The usage guide describes an offer-only setting for concurrent lines of work: the handoff stays open until explicitly accepted. That distinction matters when several sessions share one project and the next session is not necessarily the intended recipient.
Before using
What to review
The support matrix lists Linux, macOS and Windows through WSL2 as supported, while native Windows remains experimental. Some clients have lifecycle capture and MCP; others have only one integration surface. Check the exact client and platform instructions.
Capture, search and handoffs can work without an LLM provider. Optional cloud embeddings, consolidation or reranking can send stored pages, captured text or queries to the selected provider. Review capture, sharing, provider and retention settings before connecting real project sessions.
The security documentation describes sanitization as best-effort and recovered content as untrusted history. It also describes the data directory as a plain filesystem tree without added encryption at rest. Stored notes need their own access and backup choices, and important claims still need checking against current project instructions and code.
Reader fit
Who may find it relevant
Developers who revisit projects across sessions or switch between supported coding tools are the clearest audience. It suits people comfortable operating a service and maintaining captured project knowledge. Someone who needs only a small manual note may find the hooks, server and integration settings more work than that task needs.
Editorial note
Why LifeHubber lists it
LifeHubber lists ai-memory because its session continuity also leaves a readable project record. A developer revisiting an old decision can open the underlying Markdown, examine what was retained and correct the note when the project changes. Combining captured work with an editable wiki gives that developer a concrete way to maintain the knowledge being carried between tools. Retrieval can bring back a past account; the current project still determines whether it applies.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
More in AI Agents
Keep browsing this category
Explore more AI agent projects.
Paperclip
paperclipai/paperclip
A self-hosted server and dashboard for coordinating agent teams through companies, goals, roles, issues, heartbeats, budgets, approvals, and persistent activity records.
Harness²
google-research/harness2
A Google Research framework for recursively refining the prompts, tools, memory and other components around a fixed model, with parallel and sequential experiments, saved harnesses and a credential-free local demo.
Her Claude Code Session Detective
build-small-hackathon/her
A Claude Code JSONL trace reader with turn-linked explanations, single-session and multi-file project views, plus documented local setup. Its hosted Hugging Face Space was paused on 4 October 2026; the published code and handling notes remain available.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
