Add-ons & Tools (35 programs)

Pros: Vietnam-focused data and routing tailored to regional planning. Map visualization places stops geographically alongside chat output. Structured plan tables produce day-by-day schedules for review. Built on MCP for integration with modern assistant hosts.

Cons: Requires an MCP-capable host and a Node.js runtime to run. Output fidelity depends on underlying model and localized data. Not a substitute for official scheduling, booking, or confirmations.

Pros: Preserves localization file structure and contextual cues during automated edits. Sandboxed execution keeps processing on the user's infrastructure. Supports Slack, Feishu, and CLI for in-workflow interactions. Multi-tenant sessions and persistent memory for project isolation.

Cons: Depends on an MCP-compliant host to integrate with language models. Operational overhead from self-hosting and server maintenance. Output accuracy tied to the connected language model, needs human review.

Pros: Durable registers preserve long-term project decisions and preferences. Combines semantic similarity search with traditional full-text lookup. Single NativeAOT binary with local deployment and optional npm path. Local-first storage in SQLite for offline availability and privacy control.

Cons: Requires an MCP-compatible client for integration with assistants. Local SQLite database requires active maintenance and governance. Write-gated curation can omit transient context some sessions may need.

Pros: Spec-driven agents read documentation before acting, preserving technical constraints. Built-in 'critic' agents review actor outputs to catch errors and hallucinations. Zero-recompile MCP connection via JSON enables server integration without coding. Supports 25+ LLM providers through the GoAI SDK for provider flexibility.

Cons: Requires familiarity with declarative YAML workflow files. Certain MCP integrations require a Node.js environment. Designed for developer teams, not casual conversational use.

Pros: Deterministic AST extraction with tree-sitter avoids LLM calls. Provenance-labeled edges (EXTRACTED/INFERRED/AMBIGUOUS) enable auditing. MCP server lets AI assistants query structured subgraphs directly. Indexes multimodal inputs: code, PDFs, images, diagrams, transcripts.

Cons: Initial MCP client setup can require manual JSON path configuration. INFERRED relationships need human verification for high-stakes uses. Reflect feature depends on tracked feedback to improve retrievals.

Pros: MCP server lets agents return audible pronunciations to callers. Dictionary focused on developer jargon with 1,650+ entries. IPA transcriptions and phonetic respellings for every entry. Multiple access points: CLI, editor extension, and PWA.

Cons: Curated entries may not settle community debates over contested pronunciations. Installation uses developer-oriented tools (Homebrew, WSL, shell scripts). Browser PWA is offline-capable only after installation.

Pros: Machine-readable contracts enforce strict input/output validation. Full execution tracing provides audit trails for AI-driven decisions. Runs device-independent WASM across browser, Node.js, and native runtimes.

Cons: Authoring new capabilities commonly requires Rust expertise. Adoption requires protocol and pipeline integration work.

Pros: High-precision symbol mapping using tree-sitter. Built-in MCP server exposes structural context to AI agents. Generates call graphs and cyclomatic complexity metrics. Incremental indexing with bounded memory for large repos.

Cons: CLI-only workflow requires command-line proficiency. Requires building from source with a Go environment. Focused on local indexes; remote access needs host-side service.

Pros: Exposes internal Whale Cards to AI agents for on-demand retrieval. Integrates using the Model Context Protocol for agent connectivity. API-based authentication controls which agents access sensitive knowledge. Can return full playbooks and SOP content to authorized agents.

Cons: Generated outputs depend on the quality of stored internal documents. Requires an MCP-enabled client plus a Whale account and API key. Operates as a knowledge source, not a standalone conversational agent.

Pros: Near-instant search results reported for typical queries. Runs as a Go-based binary with macOS and Linux support. Self-hosted architecture keeps code and indexes on your infrastructure.

Cons: Requires self-hosted deployment and ongoing operational maintenance. No verified Windows support in the documented platforms. Scale and index management must be handled by the team.

Pros: Converts ChatGPT ZIP/JSON exports into organized Markdown files. Built-in full-text search powered by the Tantivy library. Single-binary distribution with no external runtime dependencies. One-command setup for Claude Desktop and Claude Code.

Cons: Requires obtaining ChatGPT export ZIP via the service's Data Controls. Depends on an MCP-compatible host application for model access. Limited to ChatGPT export format for direct import. Targeted at power users rather than casual, nontechnical users.

Pros: Unified memory across multiple AI coding tools and assistants. Hybrid BGE-M3 vectors plus jieba full-text search for semantic and keyword recall. Local sanitization removes secrets before storage, supporting privacy controls.

Cons: Requires self-hosting and infrastructure upkeep via Docker Compose. Search quality depends on chat clarity and extraction fidelity. Needs an MCP-compatible host and collector for cross-device synchronization.

Pros: Native MCP integration for use with MCP-compatible clients. Command-line server management for developer control. Open-source Go codebase, allowing community modifications. Handles multiple languages and dialects through connected LLMs.

Cons: Translation quality depends on the connected LLM's capabilities. Requires building from source with the Go toolchain. Public documentation does not state data-retention or training opt-out controls.

Pros: Native MCP tool-calling for direct integration with AI assistants. Accepts JSON and structured localization files for production workflows. Runs locally on Node.js across Windows, macOS, and Linux. Extensible backend connectors let teams choose AI providers.

Cons: Translation accuracy depends on the selected AI backend. Requires Node.js setup and developer configuration. Often needs an API key for the configured model. Focused on MCP workflows, not general-purpose translators.