Persistent memory for AI coding agents, storing learned architecture decisions, patterns, and bug fixes in a local SQLite database with full-text search, enabling agents to recall information across sessions.
Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find.
Install from
M8ven verifies MCPs across every public registry — install directly from whichever one you prefer.
process.env. You'll be asked to provide them before it can run.GITHUB_REF_NAMEGITHUB_REPOSITORYGITHUB_REPOSITORY_OWNERTHOTH_DATA_DIR— ~/.thoth Data directory for SQLite databaseTHOTH_DEDUPE_WINDOW_MINUTES— 15 Rolling deduplication windowTHOTH_EMBEDDING_BASE_URL— provider-specific Base URL for remote/local API providersTHOTH_EMBEDDING_DIMENSIONS— inferred for known models Optional embedding dimensions overrideTHOTH_EMBEDDING_MODEL— Use the exact model id shown by LM Studio for .THOTH_EMBEDDING_PROVIDER— lmstudio \THOTH_HTTP_DISABLED— true thoth-mem mcpTHOTH_HTTP_PORT— 7438 HTTP REST API portTHOTH_HYDE_BASE_URL— unset Optional HyDE provider base URLTHOTH_HYDE_ENABLED— true Enable HyDE dual-input semantic query expansionTHOTH_HYDE_MODEL— onnx-community/Qwen2.5-Coder-0.5B-Instruct HyDE generation model idTHOTH_HYDE_PROVIDER— transformers_local HyDE generation provider (transformers_local, ollama, lmstudio)THOTH_HYDE_TIMEOUT_MS— 4000 HyDE timeout before raw-query-only fallbackTHOTH_KG_LLM_BASE_URL— unset KG LLM provider base URL for remote providersTHOTH_KG_LLM_ENABLED— false Enable optional LLM KG enrichment for long observationsTHOTH_KG_LLM_MIN_CONTENT_CHARS— 12000 Minimum observation size that triggers LLM enrichmentTHOTH_KG_LLM_MODEL— onnx-community/Qwen2.5-Coder-0.5B-Instruct \THOTH_KG_LLM_PROVIDER— transformers_local \THOTH_KG_LLM_TIMEOUT_MS— 8000 KG LLM timeout before deterministic-only fallbackTHOTH_MAX_CONTENT_LENGTH— 100000 Max content length (warns, never truncates)THOTH_MAX_CONTEXT_RESULTS— 20 Max observations in context responseTHOTH_MAX_SEARCH_RESULTS— 20 Max search results returnedTHOTH_PREVIEW_LENGTH— 300 Search result preview lengthTHOTH_RETRIEVAL_EVAL_NOISE— Scale the retrieval eval with when you want hundreds or thousands of synthetic distractors. In PowerShell:[](https://m8ven.ai/mcp/eremesng-thoth-mem-93m8sm)