Engram MCP vs Inkwell — MCP Server Comparison | AllMCPs
Side-by-Side Model Context Protocol Comparison
Engram MCP vs Inkwell
In-depth architectural comparison of the Engram MCP and Inkwell MCP servers. Compare execution transports, security boundaries, tool capabilities, quality scores, and ready-to-paste client installation snippets for Claude, Cursor, Windsurf, and VS Code.
At a Glance & Executive Verdict
Engram MCP
Knowledge & Memory · Local stdio
Quality: 42/100 (Fair) | Auth: No auth required
Inkwell
Knowledge & Memory · Local stdio
Quality: 47/100 (Fair) | Auth: No auth required
Verdict Summary: Choose Engram MCP if you need specialized Knowledge & Memory tools running via a local process. Choose Inkwell if your workspace requires Knowledge & Memory integration with local subprocess execution. Both servers can be configured concurrently in your client's mcpServers manifest.
Which MCP Server Should You Choose?
Choose Engram MCP when:
You need dedicated capabilities in the Knowledge & Memory domain.
You prefer local stdio subprocess transport architecture.
Your security boundary fits: No auth required (Free / Open Source).
You have access to required keys: ENGRAM_DB, ENGRAM_EMBEDDING_URL.
Persistent semantic memory for AI agents. SQLite-backed, local-first, zero config. Semantic search via Ollama embeddings (nomic-embed-text) with keyword fallback. remember, recall, history, forget, and stats tools. Works with Claude Desktop, Cursor, and any MCP client.
Persistent memory for AI agents: Markdown entries, hybrid search, typed graph links
Engram MCP is categorized under Knowledge & Memory and uses a local stdio subprocess. In contrast, Inkwell belongs to Knowledge & Memory using local stdio subprocess. Select Engram MCP when you need capabilities focused on knowledge & memory and Inkwell when you require tools for knowledge & memory.
Compare recent memory activity against a baseline period to identify new entities, theme shifts, emotional changes, and anomalies.
consolidate
Consolidate working memories into long-term storage. Archives old memories beyond the keep threshold.
process_detect
Identify recurring conversation and behavior patterns: decision loops, planning sessions, learning moments, and conflict patterns.
forget
Delete memories matching criteria. Requires confirm: true. This is irreversible.
stats
Get memory statistics for a session: counts by role/tier, embeddings coverage, importance scores, and graph metrics.
graph
Query the entity relationship graph. Find connections, relationships, and paths between entities.
Inkwell Tools (8)
remember
Create or update an entry (upsert with duplicate detection, or version it via `supersede`). `entry_type` is required, and `part_of` (hub UUIDs) is required/optional/rejected per type per the schema's `membership` rule. Returns `size`, non-blocking atomicity `warnings` (Markdown headers, >3 paragrap…
recall
Read an entry with its graph relations (outgoing + backlinks). Returns `size` and `last_modified`. High-degree types (hubs) return `in_digest` — back-links grouped by linking type, plus members grouped by `part_of` — instead of an arbitrarily truncated list.
Browse entries sorted by title, filterable by tags, exact `entry_type`, and/or `part_of
tags
List all tags with entry counts
forget
Delete an entry (file and index). Warns when other entries still link to it.
rebuild
Rebuild search index from Markdown files; also runs `doctor
doctor
Check every entry against `schema.json`: dangling/superseded `kb://` targets, undeclared types, missing required frontmatter and template body fields, supernodes, tag/type collisions