AiHummer Einstein (memory plugin)
Einstein is the official long-term memory plugin for AiHummer. It gives an agent a durable, searchable memory that survives restarts and conversations, without ever turning that memory into a black box. The guiding rule is simple: canonical Markdown is the source of truth, the system indexes and proposes changes, and every promotion into long-term memory is reviewed by a human.
The plugin runs host-native as its own small Python service (standard library only — no heavyweight framework) and talks to the gateway over the contract. The in-gateway memory subsystem (claims, recall, the data-fence) is described on the concept page Memory (Einstein); this page covers the plugin that backs it.
Facts
| Field | Value |
|---|---|
| Version | 1.0.0 |
| Port | 8820 |
| Runtime | Python (standard library), host-native |
What it is
Einstein stores memory as human-readable Markdown — the canonical record a person can open, read and edit. On top of that record it builds the machinery an agent needs at turn time:
- Retrieval — fetch the memory relevant to the current conversation.
- Search — full-text and embedding-based lookup over stored facts.
- Embeddings — vectors for semantic recall, served over HTTP.
Because the Markdown is canonical, nothing about the indexes is precious: they can be rebuilt from the source of truth at any time, and a reviewer always reads the same text the agent reads.
How it is used
At turn time the gateway asks Einstein for the memory relevant to the current context. Recall is delivered to the model as a tool result wrapped in a data-fence, never as injected instructions — so a malicious note that found its way into memory cannot hijack the agent. New facts observed in a conversation are extracted as claims with evidence and queued for review rather than written straight into memory.
[!NOTE] The system indexes and proposes, but never silently rewrites memory. Promotion of a claim into long-term memory is a deliberate, human-reviewed step. Memory mode (auto / review / off) and retrieval mode (fulltext / embedding) are admin-configurable.
The v2 sidecar (advanced, write-flags off)
Einstein ships an optional v2 memory platform with richer capabilities: claim extraction, a review queue, a “dream”/consolidation pass, a memory graph, and contradiction detection. These are powerful, so they are shipped conservatively.
[!WARNING] Every write capability of the v2 sidecar is disabled by default. The advanced features index, analyse and propose; they do not mutate canonical memory until an operator explicitly enables the corresponding flag and a human approves the change. This preserves the core guarantee that memory is never rewritten behind your back.
Installation
Install Einstein in one click from the marketplace in the admin UI. The host-native deployer downloads the plugin, runs its install step, renders a sandboxed systemd unit and polls the service health endpoint before marking the installation ready. See Install & updates for the full lifecycle. There are no containers — Einstein runs as its own systemd service alongside the gateway.
Security and limits
- Source of truth is Markdown. Indexes and embeddings are derivable; the canonical text is what a human reviews and edits.
- No silent rewrites. Promotion is human-reviewed; v2 write flags are off by default.
- Data-fenced recall. Memory reaches the model as fenced tool output, never as instructions, which blocks indirect prompt injection.
- Secured Web UI. The plugin’s review/management UI is access-controlled.
- Host-native. Runs under systemd, not in a container.