knowns vs LangChain

Stores project tasks as markdown files in .knowns/tasks/ directory with Git-friendly format, enabling AI agents to maintain persistent memory across sessions. Tasks include acceptance criteria, implementation plans, and @doc/path/@task-N references that create a context graph. When an AI agent is assigned a task, it parses all embedded references, recursively follows links to documentation, and builds a complete context graph before implementation — solving the stateless AI problem where context must be re-explained each session.

Unique: Uses Git-tracked markdown files with @reference syntax for context linking instead of a centralized database, making the entire knowledge base human-readable, version-controlled, and portable. The reference resolution happens at read-time (when AI agent accesses a task) rather than at write-time, enabling dynamic context graphs that adapt as documentation changes.

vs alternatives: Unlike Jira or Linear which store context in proprietary databases, knowns makes task context Git-trackable and AI-readable; unlike simple markdown folders, it provides structured reference linking and recursive context resolution for AI agents.

Implements a Model Context Protocol (MCP) server that exposes the task and documentation system to AI agents via standardized protocol bindings. When an AI agent connects via MCP, it can query tasks, resolve references, and retrieve full context graphs without parsing markdown directly. The MCP server translates internal FileStore operations into MCP resource and tool endpoints, enabling seamless integration with Claude, GPT, and other MCP-compatible agents.

Unique: Implements MCP as a first-class integration point rather than an afterthought, making the entire task/doc system queryable via standard protocol. The MCP server translates FileStore operations into protocol-native endpoints, enabling AI agents to resolve context graphs without understanding knowns' internal markdown structure.

vs alternatives: Provides standardized MCP integration vs. custom API endpoints; enables any MCP-compatible agent to access context without custom adapters; follows protocol standards for interoperability.

Implements knowns as a TypeScript codebase that compiles to JavaScript and runs on Node.js, Deno, and browser runtimes. The build system uses Vite for bundling and supports multiple entry points (CLI, server, web UI). Core logic is runtime-agnostic, with platform-specific adapters for file I/O, HTTP, and other system operations. This enables the same codebase to run as a CLI tool, HTTP server, web application, and embedded library.

Unique: Implements a single TypeScript codebase with runtime-agnostic core logic and platform-specific adapters, enabling deployment as CLI, server, and web application without code duplication. Vite-based build system supports multiple entry points and targets.

vs alternatives: More flexible than single-runtime tools (CLI-only or server-only); enables code reuse across platforms; simpler than maintaining separate implementations for each runtime.

Provides a React-based web interface that renders the same task and documentation data as the CLI. The web UI includes a Kanban board for visual task management, a documentation browser for exploring linked docs, and a task detail view with full context. The UI communicates with the knowns server via HTTP API and WebSocket for real-time updates. All UI state is derived from the FileStore, ensuring consistency with CLI and other interfaces.

Unique: Implements web UI as a separate React application that communicates with knowns server via standard HTTP API and WebSocket, rather than embedding UI logic in the server. This enables independent UI updates and scaling.

vs alternatives: Lighter than Jira/Linear UI (no complex state management) but more polished than plain CLI; provides visual overview for non-technical stakeholders while maintaining CLI-first developer experience.

Parses @doc/path and @task-N reference syntax embedded in task descriptions and documentation, then recursively resolves all linked documents to build a complete context graph. When an AI agent requests a task, the system traverses the reference tree, fetches all linked documentation, and returns a flattened context structure. This enables AI agents to understand not just the immediate task but all architectural decisions, patterns, and related work that inform implementation.

Unique: Uses a simple @reference syntax embedded directly in markdown rather than a separate link database, making references human-readable and editable. Resolution happens at read-time with recursive traversal, enabling dynamic context graphs that adapt as documentation changes without requiring index updates.

vs alternatives: Simpler than graph database approaches (no schema, no query language) but more powerful than flat document lists; enables AI agents to discover context through reference chains rather than requiring explicit context specification.

Provides a command-line interface (knowns/kn commands) for creating, updating, and organizing tasks and documentation with built-in Kanban board state management. Tasks move through predefined states (backlog, in-progress, review, done) tracked in markdown frontmatter. The CLI supports batch operations, filtering, and status transitions. A companion web UI (React-based) renders the same data as a visual Kanban board, with both interfaces operating on the shared .knowns/ file store.

Unique: Implements a dual-interface design where CLI and web UI operate on the same file-based storage, avoiding database synchronization issues. Kanban state is stored in markdown frontmatter, making workflow status Git-trackable and mergeable.

vs alternatives: Lighter than Jira/Linear (no server, no database) but more structured than plain markdown folders; CLI-first design appeals to developers while web UI provides visual overview for non-technical stakeholders.

Maintains a version history of all task and documentation changes using a VersionStore layer that tracks file mutations over time. Each change is recorded with timestamp and metadata, enabling rollback to previous states. The versioning system operates transparently on top of the FileStore, capturing all mutations whether they come from CLI, web UI, or API calls. This enables audit trails and recovery from accidental deletions or edits.

Unique: Implements versioning at the FileStore layer (below CLI/web UI) rather than as a separate feature, capturing all mutations regardless of interface. Version history is stored alongside data files, making it portable and Git-compatible.

vs alternatives: Provides version history without relying on Git commits; enables rollback without understanding Git; simpler than full Git integration but less powerful than Git's branching model.

Stores project documentation as markdown files in .knowns/docs/ with YAML frontmatter for metadata (title, tags, created, updated). Documentation supports standard markdown syntax plus knowns-specific reference syntax (@doc/path, @task-N) for linking to other docs and tasks. The system treats documentation as first-class entities that can be queried, linked, and versioned alongside tasks. A documentation browser in the web UI enables visual navigation of the doc structure.

Unique: Treats documentation as first-class entities with structured metadata and reference linking, rather than as unstructured markdown files. Documentation is queryable, linkable, and versionable alongside tasks, creating a unified knowledge system.

vs alternatives: Simpler than wiki systems (no database, no special syntax) but more structured than plain markdown folders; enables AI agents to discover and link documentation through reference chains.

LangChain provides a Chain abstraction that sequences LLM calls, prompt templates, and tool invocations into directed acyclic graphs (DAGs). Chains support sequential execution (SequentialChain), conditional branching (RouterChain), and parallel execution patterns. The framework uses a Runnable interface that standardizes input/output contracts across all chain components, enabling composition via pipe operators and method chaining. This allows developers to build complex multi-step workflows without managing state manually.

Unique: Uses a unified Runnable interface across all components (LLMs, tools, retrievers, parsers) enabling composability via pipe operators, unlike frameworks that require separate orchestration layers for different component types. Supports both sync and async execution with identical code paths.

vs alternatives: More flexible than simple prompt chaining (like OpenAI's function calling alone) because it abstracts orchestration logic, making chains reusable and testable; simpler than full workflow engines (Airflow, Prefect) because it's optimized for LLM-specific patterns rather than general data pipelines.

LangChain's PromptTemplate class provides structured prompt engineering with variable placeholders, automatic validation, and support for few-shot learning patterns. Templates use Jinja2-style syntax for variable substitution and support dynamic example selection via ExampleSelector. The framework includes specialized templates (ChatPromptTemplate for multi-turn conversations, FewShotPromptTemplate for in-context learning) that handle formatting differences across LLM types. This enables prompt reusability, version control, and systematic experimentation without string concatenation.

Unique: Provides first-class abstractions for few-shot learning (FewShotPromptTemplate) with pluggable ExampleSelector strategies, enabling dynamic example selection based on input similarity without requiring developers to implement selection logic. Separates system prompts, conversation history, and user input in ChatPromptTemplate, making multi-turn conversations composable.

vs alternatives: More structured than manual string formatting because it validates variable names and supports semantic example selection; more specialized than generic templating engines (Jinja2) because it understands LLM-specific patterns like chat message roles and few-shot formatting.

LangChain abstracts function calling across LLM providers by converting Python functions or Pydantic models into provider-specific schemas (OpenAI function_call, Anthropic tool_use, etc.). The framework automatically generates schemas, handles argument parsing, and routes calls to the correct provider. Developers define functions once and LangChain handles provider-specific formatting. This enables tool use without learning each provider's function calling API.

Unique: Automatically converts Python functions and Pydantic models into provider-specific function calling schemas (OpenAI, Anthropic, Cohere, etc.) and handles parsing and routing transparently. Developers define tools once and LangChain handles provider-specific formatting and execution.

vs alternatives: More portable than using provider SDKs directly because function definitions are provider-agnostic; more automated than manual schema management because schemas are generated from function signatures.

LangChain supports streaming LLM output at token granularity, enabling real-time user feedback as tokens are generated. The framework provides streaming iterators and async generators that yield tokens as they arrive from the LLM. Streaming is integrated into chains and agents, so developers can stream output from complex workflows without special handling. This enables responsive user experiences where output appears in real-time rather than waiting for full completion.

Unique: Integrates streaming at the framework level so chains and agents can stream output transparently without special handling. Provides both sync and async streaming iterators and handles provider-specific streaming formats uniformly.

vs alternatives: More integrated than provider-specific streaming APIs because streaming works across chains and agents; more responsive than buffering full output because tokens appear in real-time.

LangChain provides async/await support throughout the framework, enabling concurrent execution of LLM calls, chains, and agents. All major components (LLMs, chains, retrievers, agents) have async variants (e.g., arun() alongside run()). The framework uses asyncio for Python and native async/await for Node.js. This enables high-concurrency applications that can handle multiple requests simultaneously without blocking. Async execution is transparent; developers write the same code as sync but use async/await syntax.

Unique: Provides async/await support throughout the framework with parallel async implementations of all major components. Enables transparent concurrent execution without requiring developers to manage thread pools or explicit parallelization.

vs alternatives: More integrated than manual async management because async is built into the framework; more scalable than sync-only implementations because it enables handling multiple concurrent requests.

LangChain abstracts LLM APIs behind a common BaseLanguageModel interface, supporting OpenAI, Anthropic, Cohere, Hugging Face, Ollama, and 20+ other providers. The abstraction handles provider-specific details: token counting, streaming, function calling schemas, and cost tracking. Developers write LLM-agnostic code and swap providers via configuration. The framework includes built-in retry logic, rate limiting, and fallback chains for reliability. This enables portability and cost optimization without rewriting application logic.

Unique: Implements a unified BaseLanguageModel interface that abstracts away provider differences in token counting, streaming protocols, and function calling schemas. Includes built-in retry policies, rate limiting, and cost tracking at the framework level rather than requiring developers to implement these separately for each provider.

vs alternatives: More portable than using provider SDKs directly because swapping providers requires only configuration changes; more comprehensive than simple wrapper libraries because it handles streaming, retries, and cost tracking uniformly across 20+ providers.

LangChain provides a Retriever abstraction that enables RAG by connecting LLMs to external knowledge sources. The framework supports multiple retrieval strategies: vector similarity search (via VectorStore), BM25 keyword search, hybrid search, and custom retrievers. Documents are chunked, embedded, and stored in vector databases (Pinecone, Weaviate, Chroma, FAISS, etc.). The RetrievalQA chain automatically retrieves relevant documents and passes them as context to the LLM. This enables LLMs to answer questions grounded in custom data without fine-tuning.

Unique: Provides a unified Retriever interface that abstracts different retrieval strategies (vector, keyword, hybrid, custom) and integrates seamlessly with LLM chains via RetrievalQA. Includes built-in document loaders for 50+ formats (PDF, HTML, Markdown, code files) and automatic chunking strategies, reducing boilerplate for document ingestion.

vs alternatives: More integrated than building RAG from scratch because document loading, chunking, embedding, and retrieval are unified in one framework; more flexible than specialized RAG platforms (Pinecone, Weaviate) because it supports multiple vector stores and custom retrieval logic.

LangChain's Agent abstraction enables autonomous task execution by combining LLMs with tools (functions, APIs, retrievers). The agent uses an action-observation loop: the LLM decides which tool to call based on the task, executes the tool, observes the result, and repeats until the task is complete. Agents support multiple reasoning strategies: ReAct (reasoning + acting), chain-of-thought, and tool-use patterns. The framework handles tool schema generation, argument parsing, and error recovery. This enables building autonomous systems that can decompose complex tasks without explicit step-by-step instructions.

Unique: Implements a generalized Agent interface that supports multiple reasoning strategies (ReAct, chain-of-thought, tool-use) and automatically handles tool schema generation, argument parsing, and error recovery. The action-observation loop is abstracted, allowing developers to focus on defining tools rather than implementing agent logic.

vs alternatives: More flexible than simple function calling (OpenAI's tool_choice) because it implements multi-step reasoning and tool sequencing; more accessible than building agents from scratch because it handles schema generation, parsing, and error recovery automatically.