← Back to directory
COMPARE UP TO 4 SERVERS

Compare MCP servers

Compare scores, permissions, risks, and fit in one decision-focused table.

DimensionPaperGraph MCPRead math papers with evidence, not guesses.CompartmentEncrypted, fully offline long-term memory for AI agents
FMRS55 / 100 · C75 / 100 · B
Reliability8 / 2012 / 20
Security and permissions12 / 2016 / 20
Maintenance10 / 2014 / 20
Documentation14 / 2018 / 20
Setup experience11 / 2015 / 20
Best for
  • Math-paper readers who need a strict separation between evidence and guesses
  • Researchers who want an AI agent to help map theorems and dependencies inside a local workspace
  • People who want reading artifacts as Markdown they can keep in Git or notes
  • Users who want to review external references before importing them
  • Individuals or teams who prioritise privacy and offline operation and do not want data leaving the machine
  • Users who share memory across multiple MCP clients such as Claude Code, Claude Desktop, Hermes Agent, OpenClaw, Cursor, Codex and Gemini CLI
  • Scenarios that need encrypted storage, a hash-chained audit log and verifiable vault migration
  • Users who want low-latency recall, a bundled embedding model and no extra LLM calls inside the memory layer
Not for
  • People expecting the tool to prove or verify mathematical correctness
  • Anyone wanting automatic infinite crawling of the citation chain (A cites B, B cites C)
  • Users needing semantic theorem matching or equivalence judgments for similarly worded results
  • Users with scanned or OCR-heavy PDFs where extraction may be sparse
  • Users who expect a full TeX engine for complex projects
  • Users who need cloud multi-device sync or a hosted team memory service; moving between machines means locking the vault and copying the file
  • Users who expect the server itself to run an LLM that extracts facts and decides what to remember; Compartment explicitly keeps no LLM inside and leaves that to the host model
  • Users who cannot keep a passphrase safe or who need a recovery path if it is lost, since Compartment never generates a password, seed or recovery phrase
  • Deployments that require network transports such as SSE or streamable-http; this server is stdio only and opens no ports
Required permissions
  • Read and write SQLite workspace files at user-chosen local paths
  • Read user-supplied local LaTeX projects and local PDF files
  • Download e-prints from arXiv's fixed e-print endpoint
  • Query public metadata services (Crossref, OpenAlex, arXiv) when the user runs a reference search
  • Write generated Markdown reports and manifest files into a user-specified artifact directory
  • Read and write the vault file under the user's home directory, by default ~/.compartment/memory.vault, and its sibling settings file
  • Access the session directory holding the unlock credential (configurable with COMPARTMENT_SESSION_DIR), which also carries the shared embedding process's Unix socket
  • Optionally use the macOS keychain (compartment unlock --keychain is an explicit opt-in, and it survives reboots)
  • Write or merge an mcpServers entry into each MCP client's own configuration file, taking a byte-exact backup first
  • Install a PostToolUse hook for Claude Code that captures memory files the agent writes
  • Serve a read-only dashboard on 127.0.0.1 behind a one-time random URL token
Risks and side effects
  • The workspace stores text, source spans, and proof evidence extracted from papers, so committing databases or private manuscripts can leak content
  • The Scholarly Reference Resolver only searches public metadata and stops with explicit boundaries for ambiguous, old, paywalled, or metadata-only literature, requiring manual sourcing
  • An empty dependency result only means no resolvable theorem-label references were found under the current rule; it is not evidence that the theorem has no mathematical dependencies
  • Scanned PDFs may yield sparse text and missing evidence
  • The parser is not a full TeX engine, so complex projects may need an explicit main_file
  • A leaked passphrase means the encrypted content can be decrypted; Compartment generates no recovery phrase and cannot recover a lost passphrase
  • Enabling the memory_unlock tool places the passphrase in the model's context, which is why it is off by default
  • If a 2FA keyfile is lost, for example a USB stick, the vault cannot be opened even with the passphrase
  • `export --plaintext` writes the vault as unencrypted JSONL, so that file must be handled carefully
  • Memory content can come from untrusted sources; recall wraps memories with a notice that they are stored data and content can be marked quarantined, but the host agent must still treat memory as data to limit prompt-injection risk
  • The capture hook and client integration write into the user's own settings files, although the implementation backs up and merges rather than replacing
Supported clientsMCP clients accepting JSON-style stdio configurationClaude Desktop, Claude Code, Hermes Agent, OpenClaw, Cursor, VS Code, Codex, Gemini CLI, Cline, Roo Code, Zed, OpenCode, LM Studio, AnythingLLM, BoltAI, Goose, Kiro
Tools5414