Every one of the 129 reports, by slug. Generated from each report's own frontmatter, so this list cannot drift from what the atlas actually holds — unlike the comparative report's verdict section, which is hand-written.
Looking for something else? The capability index groups systems by the seven rubric mechanisms, the comparative matrix puts eleven columns side by side, and the pattern library starts from the mechanism rather than the system.
A
a-memA-MEM — Evolving Zettelkasten memory ·agiresearch/A-memacontextAcontext — Memory as skill files ·memodb-io/Acontext· Scope enforced, Human reviewadk-pythonGoogle ADK — A contract with no delete ·google/adk-python· Scope enforcedagent-afkagent-afk — A citation or an [unverified] tag ·griffinwork40/agent-afk· Negative evalsagent-frameworkMicrosoft Agent Framework — The contract AutoGen became ·microsoft/agent-framework· Scope enforcedagent-memory-supabaseAgent Memory on Supabase — One table, argued for ·reescalder/agent-memory-supabase· Bi-temporal, Scope enforcedagentmemoryagentmemory — Hook-native coding-agent memory ·rohitg00/agentmemory· Scope enforcedagentrecall-xAgentRecall-X — A correction that loses its veto ·Goldentrii/AgentRecall-X· Trust state, Scope enforcedagentswarmsAgentSwarms — Keyword memory in a trigger ·AgentSwarms-fyi/agentswarms· Scope enforced, Human reviewagnoAgno — Framework-native learning stores ·agno-agi/agno· Scope enforced, Human review, Negative evalsai-memoryai-memory — Cross-harness handoff ·akitaonrails/ai-memory· Scope enforcedatomic-agentAtomic Agent — Evaluated memory fabric ·AtomicBot-ai/atomic-agent· Bi-temporalaukora-kernelAukora Kernel — Receipted memory ·aumara-xyz/aukora-kernel· Scope enforced, Mutation audit, Negative evalsauraAura — Receipts that chain, beliefs that don't survive ·youngbryan97/aura· Trust state, Mutation auditautogenAutoGen — Memory without identity ·microsoft/autogen
B
basic-memoryBasic Memory — Human-editable Markdown graph ·basicmachines-co/basic-memory· Scope enforcedbuzzBuzz — Memory as an encrypted protocol ·block/buzz· Scope enforcedbyteroverByteRover — Curated coding-agent memory ·campfirein/cipher
C
camelCAMEL — Message store with vector recall ·camel-ai/camelclaude-memClaude-Mem — Hook-native coding-agent memory ·thedotmack/claude-mem· Scope enforcedclioCLIO — Corroboration tiers in pure Perl ·SyntheticAutonomicMind/CLIO· Trust state, Human reviewcogneeCognee — Knowledge-graph memory control plane ·topoteretes/cognee· Scope enforcedcore-memoryCore Memory — Grounding-gated trust ·JohnnyFiv3r/Core-Memory· Trust state, Bi-temporal, Scope enforced, Human reviewcortexCortex — Approval on the read, not the write ·CortexPrism/cortex· Scope enforced, Human reviewcosmonapseCosmonapse — A memory contract with a failure vocabulary ·Cosmonapse/cosmonapse-corecowagentCowAgent — Daily-distilled memory ·zhayujie/CowAgent· Scope enforcedcrewaiCrewAI — Hierarchical scope paths ·crewAIInc/crewAI· Scope enforced, Negative evalscsmCSM — Deterministic continuity runtime ·NovasPlace/CSM· Scope enforced, Mutation audit, Negative evalsctxctx — Guarded consolidation ·ActiveMemory/ctx· Scope enforced, Mutation audit
D
daimonDaimon — Session-boundary checkpoint memory ·Daily-Nerd/daimon· Tombstone, Trust state, Scope enforced, Mutation audit, Human review, Negative evalsdextoDexto — CRUD, and no retrieval on purpose ·truffle-ai/dexto· Human review
E
eccECC — A vault that stays unreviewed ·affaan-m/ECC· Scope enforcedelastic-atlasElastic Atlas — Evaluated demo ·noamschwartz/atlas-memory-demo· Scope enforcedempryoEmpryo — Retrieval that reads the repo ·proxysoul/Empryo· Scope enforced, Human reviewengramEngram — Coding-agent local memory ·Gentleman-Programming/engram· Scope enforced, Human revieweverosEverOS — Cases, skills and a cascade ·EverMind-AI/EverOS· Scope enforced, Negative evals
G
generative-agentsGenerative Agents — Observation-reflection ancestor ·joonspk-research/generative_agentsgenericagentGenericAgent — Memory as written policy ·lsdefine/GenericAgentgini-agentGini Agent — Reimplemented memory model ·Open-Curiosity/gini-agent· Trust state, Bi-temporal, Scope enforcedgitlordGitLord — Git as the agent's event log ·yashneil75/gitlordgobiiGobii — Memory as a database the agent writes ·gobii-ai/gobii-platform· Scope enforcedgoodai-ltmGoodAI LTM — The lifecycle the frameworks dropped ·GoodAI/goodai-ltmgraphifyGraphify — Work memory beside a code graph ·Graphify-Labs/graphify· Trust state, Negative evalsgraphitiGraphiti — Bi-temporal context graph ·getzep/graphiti· Bi-temporal, Scope enforced
H
helmHelm — Confidence-ratcheted SQLite memory ·GOODMAN-PRO/helm· Negative evalshermes-agentHermes Agent — Prompt-cache-bounded memory ·NousResearch/hermes-agent· Human reviewhindsightHindsight — Retain, recall, and reflect ·vectorize-io/hindsight· Scope enforcedhipporagHippoRAG — Associative graph memory ·OSU-NLP-Group/HippoRAGholographicHolographic — Vector-symbolic memory ·NousResearch/hermes-agenthonchoHoncho — Peer and session representation ·plastic-labs/honcho· Scope enforced
J
jugglerJuggler — A gitignored notebook with a delete button ·juggler-ai/juggler· Human review
L
langmemLangMem — Library primitives ·langchain-ai/langmem· Scope enforcedletheLethe — Verified forgetting, signed ·deeplethe/lethe· Mutation audit, Negative evalslettaLetta — Agent-runtime memory ·letta-ai/letta· Scope enforcedlivingfeedLivingFeed — Forgetting as a feature, in Korean ·showjihyun/livingfeed· Scope enforcedllamaindexLlamaIndex — Composable memory blocks ·run-llama/llama_indexllm-wiki-memoryllm-wiki-memory — Filesystem and git memory ·ctxr-dev/llm-wiki-memory· Scope enforced, Human reviewlogseqLogseq — A human knowledge base agents can write to ·logseq/logseqloongflowLoongFlow — Stochastic recall ·baidu-baige/LoongFlowlorekitLoreKit — Multi-tenant lesson store ·mthines/lorekit· Scope enforced, Mutation audit, Human review
M
magic-contextMagic Context — Verified-against-code memory ·cortexkit/magic-context· Trust state, Scope enforced, Mutation auditmastra-observational-memoryMastra Observational Memory — Observer–reflector context ·mastra-ai/mastra· Scope enforcedmateclawMateClaw — Scoped provider SPI ·mateaix/mateclaw· Scope enforcedmem0Mem0 — Library-first memory ·mem0ai/mem0· Scope enforced, Mutation auditmem0sharpMem0Sharp — Mem0's architecture, rebuilt in C# ·jihadkhawaja/mem0sharp· Scope enforced, Mutation auditmemantoMemanto — Resolved conflict ·moorcheh-ai/memanto· Scope enforced, Human reviewmemaryMemary — Knowledge-graph agent memory ·kingjulio8238/MemarymembaseMembase — Signature-owned conversation memory ·unibaseio/membase· Scope enforcedmementoMemento — Memory sealed until a date ·xD4O/memento· Scope enforcedmemex-zero-ragMeMex Zero-RAG — Enforcement by instruction ·JPeetz/MeMex-Zero-RAGmemledgerMemLedger — Provenance that answers, and does not enforce ·riktar/memledger· Trust state, Mutation auditmemmachineMemMachine — Episode-preserving dual memory ·MemMachine/MemMachine· Scope enforcedmemobaseMemobase — Structured user profiles ·memodb-io/memobase· Scope enforcedmemoraMemora — Rehearsable correction ·agentic-box/memora· Mutation audit, Human reviewmemoriMemori — Portable store, hosted judgement ·MemoriLabs/Memori· Scope enforcedmemory-engineMemory Engine — Agents as principals ·timescale/memory-engine· Bi-temporal, Scope enforcedmemoryosMemoryOS — Heat-promoted tiers ·BAI-LAB/MemoryOSmemosMemOS — Memory operating system ·MemTensor/MemOS· Scope enforcedmempalaceMemPalace — Verbatim evidence memory ·MemPalace/mempalace· Scope enforcedmemumemU — Documented decisions ·NevaMind-AI/memUmemvidMemvid — Rewindable memory ·memvid/memvid· Bi-temporal, Mutation auditmercury-agentMercury Agent — Graded personal memory ·cosmicstack-labs/mercury-agent· Human reviewmetaclawMetaClaw — Self-tuning memory policy ·aiming-lab/MetaClaw· Scope enforcedminecontextMineContext — Prospective memory, inferred ·volcengine/MineContextmirixMIRIX — Multi-agent typed memory ·Mirix-AI/MIRIX· Scope enforced, Negative evalsmnemopiMnemopi — A decay curve per memory type ·can1357/oh-my-pi· Negative evalsmoltisMoltis — File-corpus memory ·moltis-org/moltis
N
nanobotnanobot — Dual-cursor file memory ·HKUDS/nanobotnekoProject N.E.K.O. — Memory that must not reopen a wound ·Project-N-E-K-O/N.E.K.O· Scope enforced, Mutation audit, Negative evalsnemoclawNemoClaw — Memory as an operand ·NVIDIA/NemoClawneo4j-agent-memoryNeo4j Agent Memory — Memory of its own reasoning ·neo4j-labs/agent-memory· Bi-temporal, Scope enforcednooa-memoryNOOA Memory — Cognitive-model recall ·NVIDIA-NeMo/labs-OO-Agents· Scope enforcednpcpynpcpy — Memory that waits for a yes ·npc-worldwide/npcpy· Trust state, Human review
O
open-coworkopen-cowork — Evaluated memory ·OpenCoworkAI/open-cowork· Negative evalsopenclawOpenClaw — Host-runtime plugin memory ·openclaw/openclaw· Scope enforcedopencodeOpenCode — Hooks, no contract ·anomalyco/opencodeopenhumanOpenHuman — Taint-gated memory ·tinyhumansai/openhuman· Scope enforcedopenvikingOpenViking — Filesystem context database ·volcengine/OpenViking· Scope enforcedopenworkerOpenWorker — Policy over machinery ·andrewyng/openworker· Scope enforcedopenyakOpenYak — One document per workspace ·openyak/openyak· Scope enforced, Human reviewoptmemOptMem — No background work ·VictorTaelin/OptMem· Mutation audit
P
palazzoPalazzo — An audit log that can refuse a delete ·calibrae/palazzo· Mutation auditpiPi — Branchable session substrate ·earendil-works/pipowermemPowerMem — Decay on the read path ·oceanbase/powermem· Scope enforcedproject-golemProject Golem — A list of what the owner said no to ·Arvincreator/project-golempydantic-ai-harnessPydantic AI Harness — Markdown notebook with database discipline ·pydantic/pydantic-ai-harness· Scope enforced, Negative evals
Q
qwen-codeQwen Code — Team memory in git ·QwenLM/qwen-code· Scope enforced, Human review
R
rainboxRainBox — Operator-governed memory ·neoneye/RainBox· Tombstone, Trust state, Scope enforced, Human reviewredis-agent-memory-serverRedis Agent Memory Server — Working/long-term memory service ·redis/agent-memory-server· Scope enforcedremeReMe — Markdown memory that publishes its bad scores ·agentscope-ai/ReMerisuaiRisuAI — Three memory systems, one repository ·kwaroran/RisuAI· Human review
S
second-meSecond Me — The atlas's only parametric memory ·mindverse/Second-Me· Human reviewsillytavernSillyTavern — Memory people write by hand ·SillyTavern/SillyTavernsimplememSimpleMem — Compression-first research memory ·aiming-lab/SimpleMem· Scope enforced, Mutation auditskalesSkales — Ask it to forget ·skalesapp/skalessoul-of-waifuSoul of Waifu — Rollback that nothing calls ·jofizcd/Soul-of-Waifu· Mutation auditsupermemorySupermemory — Hosted memory product ·supermemoryai/supermemory· Scope enforcedswafraSwafra — Compact local graph-RAG ·kunal12203/swafra
T
tencentdb-agent-memoryTencentDB Agent Memory — Layered symbolic memory ·TencentCloud/tencentdb-agent-memoryterse-memoryTERSE Memory — A linter where the verbs should be ·terse-lang/terse· Trust statetigrimosrTigrimOSR — Proposed skills, staged on disk ·Sompote/TigrimOSR· Trust state, Scope enforced, Human reviewtokenmizerTokenMizer — A status for not knowing ·Shweta-Mishra-ai/tokenmizer· Trust state, Mutation audit
U
universal-memory-engineUniversal Memory Engine — A suppression the write gate obeys ·12ziyad/universal-memory-engine· Tombstone, Trust state, Scope enforced, Mutation audit, Human review, Negative evals
V
verelVerel — Verification-first memory ·amitpatole/verel· Tombstone, Trust state, Bi-temporal, Scope enforced, Mutation audit, Human review, Negative evalsvirtualwifeVirtualWife — normalize_scores does not normalize ·yakami129/VirtualWifevoyagerVoyager — Procedural skill memory ·MineDojo/Voyager
W
waku-agentWaku Agent — Gated memory ·ShenSeanChen/waku-agent