NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Go modules · #2353 by repository stars
Last release 5 days ago
03 Oct 2026
Release timing varies
gaps range from 8 days to 2 months
Most releases are documented
notes for 26 of 29 stable releases
Nothing withdrawn
no release was ever pulled
7 months old
39 releases · first in 2026
One column per month.
v0.2.10 improves Memory retrieval performance, storage efficiency, and OpenCode/Windows integrations. Memory and the local Agency authority remain ava
v0.2.10 improves Memory retrieval performance, storage efficiency, and OpenCode/Windows integrations. Memory and the local Agency authority remain available in one mnemon binary.
MNEMON_EMBED_KEEP_ALIVE.mnemon gc --compact for database compaction, including explicit handling of busy checkpoints.The committed benchmark compares the optimized Memory implementation with the preceding master revision, using five alternating runs at N=1/10/100/1,000/10,000. At N=10,000, median hybrid-recall CPU time improved from 93.06 ms to 66.53 ms and cumulative allocation fell by 37.7%; atomic batch-insert CPU time improved from 524.33 ms to 431.60 ms. These are synthetic local measurements, with small-input tradeoffs documented alongside the gains. Exact vector search remains exact.
See the full analysis and raw samples and 中文报告.
npm install --global @mnemon-dev/mnemon@0.2.10
mnemon --versionExisting npm-managed installations can also run mnemon update. Native archives are available below for macOS, Linux, and Windows on amd64 and arm64, with SHA-256 checksums. Homebrew users can upgrade through the mnemon-dev/tap tap.
To refresh existing OpenCode or ZCode integrations, rerun mnemon setup for that target with the original project/global scope, then restart the host application. Updating the binary alone does not replace previously installed hooks.
Generate Homebrew casks with postflight_steps to eliminate the deprecated postflight warning.
v0.2.9 improves memory corrections, multilingual recall, and installation reliability. Memory and the local Agency authority remain available in one mnemon binary.
supersedes links so a correction can demote an older insight during recall while preserving its history. Support the relation in import, related queries, and brief recall output.postflight_steps to eliminate the deprecated postflight warning.npm install --global @mnemon-dev/mnemon@0.2.9
mnemon --versionExisting npm-managed installations can also run mnemon update.
Native archives are available below for macOS, Linux, and Windows on amd64 and arm64, with SHA-256 checksums. Homebrew users can upgrade through the mnemon-dev/tap tap.
v0.2.8 ships Mnemon Memory and the local Agency authority in one mnemon binary.
v0.2.7 fixes --readonly recall on Windows and when using a relative --data-dir path.
v0.2.7 fixes --readonly recall on Windows and when using a relative --data-dir path.
#, and %. This resolves the misleading SQL logic error: out of memory (1) failure in #123.The fix is included in PR #125, with native Windows and Linux CI coverage.
--readonly still assumes an immutable, static snapshot. It should not be used to follow a database that another process is actively changing.
Archives are available below for macOS, Linux, and Windows, each on amd64 and arm64. Verify the download against checksums.txt. Existing Homebrew Cask users can update with brew upgrade --cask mnemon.
aab97af8002bf0c8ac18d88da64293e41c9eb121.--readonly recall now resolves and encodes SQLite file URI paths correctly
on Windows and when --data-dir is relative to the current directory. This
fixes the misleading SQL logic error: out of memory (1) failure reported
in #123.mode=ro and immutable=1: they reject database
writes, preserve recall counters and oplog, and create no WAL/SHM sidecars.#, and %.v0.2.6 ships Mnemon Memory and the local Agency authority in one mnemon binary.
v0.2.6 ships Mnemon Memory and the local Agency authority in one mnemon binary.
Nothing published for this version
v0.2.5 ships Mnemon Memory and the local Agency authority in one mnemon binary.
v0.2.4 ships Mnemon Memory and the local Agency authority in one mnemon binary.
v0.2.3 ships Mnemon Memory and the local Agency authority in one mnemon binary.
v0.2.3 ships Mnemon Memory and the local Agency authority in one mnemon binary.
v0.2.2 ships Mnemon Memory and the local Agency authority in one mnemon binary.
Nothing published for this version
v0.2.1 — Memory Store Reliability
v0.2.1 strengthens Mnemon Memory's SQLite maintenance paths without changing the CLI or public command structure.
prune oplog entry for every soft-deleted insight, including the active count and configured capacity.edges table migration so schema detection works with foreign-key enforcement enabled.Existing stores already using the current four-edge schema are unchanged. For legacy stores whose constraint still permits the obsolete narrative edge type, the migration applies the existing four-graph policy: narrative edges are removed while all other edges are retained.
Release artifacts are available for macOS and Linux on amd64 and arm64.
brew install --cask mnemon-dev/tap/mnemon
# or
go install github.com/mnemon-dev/mnemon@v0.2.1Full changelog: v0.2.0...v0.2.1
v0.2.1 聚焦 Memory SQLite 存储可靠性：自动裁剪现在会逐条写入 oplog；旧版 edges 表迁移可在启用外键及存在孤立边的情况下安全完成，并避免探针残留或误删真实数据。本版本同时整理了 Memory / Agency 的内部包边界，不改变 CLI、协议或状态路径。
Nothing published for this version
v0.2.0 integrates Mnemon Agency Preview into the single mnemon executable while keeping Mnemon Memory as the established root command surface.
v0.2.0 integrates Mnemon Agency Preview into the single mnemon executable while keeping Mnemon Memory as the established root command surface.
View → Intent → Receipt interaction loop.mnemon agency ....mnemond is the local daemon and protocol role behind Agency. It is not a second executable, and it does not replace an Agent's planner, tools, provider, or model runtime.
Agency is opt-in and remains separate from Memory:
mnemon root, including remember, recall, and setup;mnemon setup continues to configure Memory;mnemon agency setup separately enables Agency for one project;.mnemon/agency directory.This release also fixes a Memory safety issue: conflict candidates are no longer silently removed, and cosine-only similarity is no longer sufficient to replace an existing memory.
Agency is actively evolving and is intended for evaluation and controlled use. Its command surface, Runtime integration, protocol, peer exchange, and on-disk/wire formats may change before an explicit stability milestone. This Preview status applies to Agency; Memory retains its separate product contract.
The obsolete experimental Harness surfaces and event-emission bridge, including mnemon event and MNEMON_HARNESS_EVENT_EMIT, have been retired without compatibility shims.
Release artifacts are available for macOS and Linux on amd64 and arm64. Windows archives are not published for v0.2.0.
brew install --cask mnemon-dev/tap/mnemon
# or
go install github.com/mnemon-dev/mnemon@v0.2.0Existing Homebrew Formula installations are migrated to the v0.2.0 Cask through the tap migration record.
Full changelog: v0.1.17...v0.2.0
v0.2.0 首次把 Mnemon Agency Preview 集成到单一 mnemon 可执行文件中。Memory 的根级命令、Store、SQLite 数据和既有 setup 链路保持独立；Agency 通过 mnemon agency ... 按项目单独启用，状态保存在 .mnemon/agency。
Agency 引入 View → Intent → Receipt 循环、Handling / Artifact / Reference 三类协作对象、Pi-first Hook 接入和显式 Peer 协作。Agency 仍处于持续迭代的 Preview 阶段，其命令、Runtime 接口、协议及持久化/传输格式可能继续变化；这不改变 Mnemon Memory 的独立产品契约和主使用链路。
v0.1.17 broadens setup support for additional coding hosts, fixes hook/runtime edge cases, and advances the experimental harness runtime work.
v0.1.17 broadens setup support for additional coding hosts, fixes hook/runtime edge cases, and advances the experimental harness runtime work.
Note: harness modules, harness documentation, and harness evaluation assets in this repository remain experimental and are not part of this release's public stability guarantee.
mnemon setup.mnemon remember / mnemon link directly instead of recursively spawning another subagent.go build -o mnemon ..Full changelog: v0.1.16...v0.1.17
Nothing published for this version
v0.1.16 focuses on stable CLI, storage, and entity-graph behavior.
v0.1.16 focuses on stable CLI, storage, and entity-graph behavior.
Note: harness modules, harness documentation, and harness evaluation assets in this repository remain experimental and are not part of this release's public stability guarantee.
v0.1.15 focuses on stable CLI, storage, and entity-graph behavior.
v0.1.15 focuses on stable CLI, storage, and entity-graph behavior.
Note: harness modules, harness documentation, and harness evaluation assets in this repository remain experimental and are not part of this release's public stability guarantee.
mnemon setup --target cursor now installs a Cursor integration, including a
Mnemon skill, prompt files, and Cursor lifecycle hooks under .cursor/ or
~/.cursor/. The session-start hook primes Cursor with Mnemon guidance and
memory status, the stop hook can nudge for durable-memory writeback, and an
optional pre-compaction hook can remind users to preserve durable context.mnemon setup --eject --target cursor removes the installed Cursor skill and
Mnemon hook entries while preserving unrelated Cursor hook configuration.cursor binary or ~/.cursor/ config directory is present.mnemon setup continue to keep formal integrations under cmd/ and
internal/.v0.1.14 focuses on stable CLI, storage, and entity-graph behavior.
v0.1.14 focuses on stable CLI, storage, and entity-graph behavior.
Note: harness modules, harness documentation, and harness evaluation assets in this repository remain experimental and are not part of this release's public stability guarantee.
mnemon setup for Claude Code from $HOME no longer writes relative hook commands into the user-global ~/.claude/settings.json. Mnemon now detects this collision, resolves symlinks, honors CLAUDE_CONFIG_DIR, and writes absolute hook paths so hooks work from any future session directory.go build -o mnemon .make test (147 E2E checks)go test ./... in a clean worktreemnemon_0.1.14_darwin_amd64.tar.gz, mnemon_0.1.14_darwin_arm64.tar.gzmnemon_0.1.14_linux_amd64.tar.gz, mnemon_0.1.14_linux_arm64.tar.gzmnemon_0.1.14_windows_amd64.zipchecksums.txtfloat32 values instead of
float64, cutting embedding storage size roughly in half while keeping the
public vector API and cosine-similarity calculations in float64.float64
embedding blobs to the new float32 storage format. The migration records its
completion in SQLite PRAGMA user_version and leaves malformed or
non-legacy-looking blobs untouched rather than blocking database startup.$HOME/.claude
collision. When mnemon setup is run from $HOME, the apparent local
.claude/ directory is actually Claude Code's user-global config directory;
Mnemon now writes absolute hook commands in that case so hooks resolve
correctly from every future Claude Code session directory.CLAUDE_CONFIG_DIR-aware coverage for Claude Code
config collision detection, plus regression tests that genuine project-local
installs keep relative hook paths.float32 vector
serialization, plus legacy float64 deserialization coverage for migration.Nothing published for this version
Nothing published for this version
Nothing published for this version
English/Chinese import guide coverage and setup skill guidance for using mnemon import on historical chat exports.
mnemon import on historical chat exports.mnemon import bulk-imports memory draft JSON files into a selected Mnemon store. Imported insights use the normal write path for duplicate and conflic
mnemon import <file> bulk-imports memory draft JSON files into a selected
Mnemon store. Imported insights use the normal write path for duplicate and
conflict detection, embeddings when Ollama is available, entity/causal/
semantic graph construction, lifecycle scoring, and capacity pruning.internal/importdraft defines the public draft schema for historical chat
imports:
schema_version, source, insights, and optional edgescontent, category, importance, tags,
entities, source, and RFC 3339 created_atsource_index, target_index, edge_type,
weight, and reasonmnemon import --dry-run validates draft structure without writing to the
database, and mnemon import --no-diff inserts all draft insights without
duplicate/conflict detection.memory_draft.json.created_at is older than existing store contents. The import path disables
real-time temporal edge generation, then repairs affected source timelines
after all backdated insights are written so imported memories are inserted
into the global chronological chain.AutoPrune runs. Import finalization inserts explicit edges, repairs
temporal edges, refreshes touched insight scores, and only then applies
pruning.EngineOptions with TemporalMode,
allowing import to skip real-time temporal generation without changing the
behavior of mnemon remember.BREAKING: recall now emits a compact, LLM-friendly JSON shape by default. The default result entries keep id, content, category, importance, intent, m
recall now emits a compact, LLM-friendly JSON shape by
default. The default result entries keep id, content, category,
importance, intent, matched_via, confidence, and rounded score,
while omitting verbose debug fields such as signals, timestamps,
access_count, tags, entities, source, and traversal metadata. (#3)recall payload remains available with --verbose for
scripts, debugging, and callers that still need meta, signals, or the
full embedded insight object.recall --verbose flag to restore the previous full recall response.--embed-model global flag exposes the existing MNEMON_EMBED_MODEL env var on the CLI, applied consistently across embed, recall, and remember. Useful
--embed-model global flag exposes the existing MNEMON_EMBED_MODEL env
var on the CLI, applied consistently across embed, recall, and
remember. Useful for switching to multilingual models such as
nomic-embed-text-v2-moe:latest on code-switched corpora without exporting
the env var. Note: switching to a model with a different output dimension
silently invalidates existing embeddings — backfill after changing.Nothing published for this version
Privacy-safe memory receipts via mnemon receipt, allowing users to show memory activity with identifiers, timestamps, summaries, and relation metadata
mnemon receipt, allowing users to show
memory activity with identifiers, timestamps, summaries, and relation metadata
without exposing stored memory content.Codex integration: mnemon setup --target codex deploys the mnemon skill, prompt files, and Codex lifecycle hooks (SessionStart, UserPromptSubmit, Stop
mnemon setup --target codex deploys the mnemon skill,
prompt files, and Codex lifecycle hooks (SessionStart, UserPromptSubmit,
Stop) into .codex/ or ~/.codex/. mnemon setup --eject --target codex
removes the installed Codex surface while preserving unrelated hooks.mnemon setup --target nanobot deploys a skill file to
.nanobot/skills/mnemon/SKILL.md (local) or ~/.nanobot/workspace/skills/mnemon/SKILL.md
(global, recommended). mnemon setup --eject removes it. Detection is automatic
when the nanobot binary or ~/.nanobot/workspace/ directory is present.cmd/
and internal/.Nothing published for this version
Diff now sorts matches by Similarity descending before selecting the overall suggestion. Previously, KeywordSearch ordered candidates by token overlap
Diff now sorts matches by Similarity descending before selecting the overall
suggestion. Previously, KeywordSearch ordered candidates by token overlap score,
so a high-keyword-score candidate classified as ADD could mask a lower-keyword-score
candidate with higher Jaccard similarity that should have been UPDATE or DUPLICATE."not" from negation words — it appears in virtually all
scientific prose and caused unrelated records to be classified as CONFLICT.ContentSimilarity) to
Jaccard (|A∩B|/|A∪B|) — penalises texts that share vocabulary but
differ in most tokens, preventing formulaic records from scoring as UPDATE.remember --entity-mode to choose merge, provided, or auto entity handling.
remember --entity-mode to choose merge, provided, or auto entity handling.MNEMON_EMBED_DIMENSIONS env var for Matryoshka dimension truncation (e.g., 256-dim instead of 768)
MNEMON_EMBED_DIMENSIONS env var for Matryoshka dimension truncation (e.g., 256-dim instead of 768)Nothing published for this version
Nothing published for this version
Nothing published for this version
OpenClaw full integration: internal hook (agent:bootstrap) + plugin (before_prompt_build)
agent:bootstrap) + plugin (before_prompt_build)mnemon setup --target openclaw now auto-deploys skill, hook, plugin, and configNothing published for this version
Core CRUD: remember, recall, forget, search, status, log
Initial public release.
remember, recall, forget, search, status, logmnemon store list|create|set|remove)MNEMON_STORE environment variable and --store CLI flag for store selection~/.mnemon/mnemon.db to ~/.mnemon/data/default/nomic-embed-text)mnemon viz)mnemon setup interactive installer with --eject supportYour coding agent can read these notes before it upgrades. Set up the MCP server →