AI coding agents have two memory problems that CLAUDE.md and Rules files don't solve: (1) long conversations get compressed and context silently disappears \u2014 the agent forgets decisions made 2 hours ago in the same session, and (2) memory is locked to one tool on one machine. Switch from Claude Code to Cursor, or from your laptop to your desktop, and everything is gone.
I built hmem to fix both. It's an MCP server that gives AI agents persistent, hierarchical memory stored in a local SQLite file. The same .hmem file works across Claude Code, Cursor, Windsurf, OpenCode, and Gemini CLI \u2014 on any machine. Your agent's knowledge is portable.
The key idea is borrowed from how human memory works: you remember rough outlines first and recall details on demand. hmem has 5 depth levels. At session start, the agent loads only Level 1 summaries (~20 tokens). It drills deeper into specific memories only when needed \u2014 L2 for context, L3-L5 for raw details. Unlike a flat MEMORY.md that gets injected wholesale (3000-8000 tokens every time), hmem loads only what's relevant.
Install: `npx hmem-mcp init` (interactive setup \u2014 detects your installed tools and writes the MCP config).
This is beta software. I've been using it in production across two machines with 100+ memory entries and it's been stable, but the API surface may still change. Would appreciate feedback.
GitHub: https://github.com/Bumblebiber/hmem\nnpm: https://www.npmjs.com/package/hmem-mcp\nLicense: MIT"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: Hmem \u2013 Persistent hierarchical memory for AI coding agents (MCP)"}},"_tags":["story","author_Bumblebiber","story_47103237","show_hn"],"author":"Bumblebiber","children":[47103263,47115673,47135750],"created_at":"2026-02-21T18:19:47Z","created_at_i":1771697987,"num_comments":3,"objectID":"47103237","points":2,"story_id":47103237,"story_text":"My Claude told me to post this here :D
AI coding agents have two memory problems that CLAUDE.md and Rules files don't solve: (1) long conversations get compressed and context silently disappears \u2014 the agent forgets decisions made 2 hours ago in the same session, and (2) memory is locked to one tool on one machine. Switch from Claude Code to Cursor, or from your laptop to your desktop, and everything is gone.
I built hmem to fix both. It's an MCP server that gives AI agents persistent, hierarchical memory stored in a local SQLite file. The same .hmem file works across Claude Code, Cursor, Windsurf, OpenCode, and Gemini CLI \u2014 on any machine. Your agent's knowledge is portable.
The key idea is borrowed from how human memory works: you remember rough outlines first and recall details on demand. hmem has 5 depth levels. At session start, the agent loads only Level 1 summaries (~20 tokens). It drills deeper into specific memories only when needed \u2014 L2 for context, L3-L5 for raw details. Unlike a flat MEMORY.md that gets injected wholesale (3000-8000 tokens every time), hmem loads only what's relevant.
Install: `npx hmem-mcp init` (interactive setup \u2014 detects your installed tools and writes the MCP config).
This is beta software. I've been using it in production across two machines with 100+ memory entries and it's been stable, but the API surface may still change. Would appreciate feedback.
GitHub: https://github.com/Bumblebiber/hmem\nnpm: https://www.npmjs.com/package/hmem-mcp\nLicense: MIT","title":"Show HN: Hmem \u2013 Persistent hierarchical memory for AI coding agents (MCP)","updated_at":"2026-03-05T23:34:20Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"winci"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: Mimirs \u2013 persistent local memory for AI coding agents (MCP)"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/TheWinci/mimirs"}},"_tags":["story","author_winci","story_48491918","show_hn"],"author":"winci","children":[48491926],"created_at":"2026-06-11T15:42:48Z","created_at_i":1781192568,"num_comments":0,"objectID":"48491918","points":2,"story_id":48491918,"title":"Show HN: Mimirs \u2013 persistent local memory for AI coding agents (MCP)","updated_at":"2026-06-12T00:47:20Z","url":"https://github.com/TheWinci/mimirs"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"imyax"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Fixhive \u2013 collective fix memory for AI coding agents (MCP plugin)"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://fixhive-landing.vercel.app"}},"_tags":["story","author_imyax","story_47732767"],"author":"imyax","children":[47732768],"created_at":"2026-04-11T18:17:00Z","created_at_i":1775931420,"num_comments":0,"objectID":"47732767","points":2,"story_id":47732767,"title":"Fixhive \u2013 collective fix memory for AI coding agents (MCP plugin)","updated_at":"2026-04-11T18:52:10Z","url":"https://fixhive-landing.vercel.app"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"zawakin"},"story_text":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai","agent","mcp"],"value":"AI agents are getting more capable, but we're increasingly in the dark\nabout what they're actually doing. They run complex multi-step workflows,\ncall dozens of tools, reason through problems - and we just watch the\noutput scroll by. It's a black box, and humans end up being led around\nby the agent rather than understanding it.
I wanted to flip this. The key insight: all these agents (Claude Code,\nCodex, Gemini) already write detailed logs. The problem is they're in\ndifferent locations, different formats, incompatible schemas.
agtrace normalizes this "observation layer" across providers:
- Auto-discovers logs from Claude, Codex, Gemini\n- Converts them into a unified event timeline\n- Exposes this via CLI, TUI dashboard, and MCP
The MCP part is what makes it interesting for agents themselves. An agent\ncan now query its own past sessions:
- "What approach did we take when we refactored auth last week?"\n- "Show me errors from yesterday's session"\n- "How did we handle this edge case before?"
This enables agent self-reflection - using execution history to inform\ncurrent decisions.
Built in Rust for safety and speed. 100% local, no cloud dependencies.\nThe database is just a pointer index to original logs - rebuilable anytime.
Happy to discuss the architecture or use cases."},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: AgTrace \u2013 Observability for AI Coding Agents via MCP (Claude Code etc.)"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/lanegrid/agtrace"}},"_tags":["story","author_zawakin","story_46499885","show_hn"],"author":"zawakin","created_at":"2026-01-05T15:27:57Z","created_at_i":1767626877,"num_comments":0,"objectID":"46499885","points":2,"story_id":46499885,"story_text":"AI agents are getting more capable, but we're increasingly in the dark\nabout what they're actually doing. They run complex multi-step workflows,\ncall dozens of tools, reason through problems - and we just watch the\noutput scroll by. It's a black box, and humans end up being led around\nby the agent rather than understanding it.
I wanted to flip this. The key insight: all these agents (Claude Code,\nCodex, Gemini) already write detailed logs. The problem is they're in\ndifferent locations, different formats, incompatible schemas.
agtrace normalizes this "observation layer" across providers:
- Auto-discovers logs from Claude, Codex, Gemini\n- Converts them into a unified event timeline\n- Exposes this via CLI, TUI dashboard, and MCP
The MCP part is what makes it interesting for agents themselves. An agent\ncan now query its own past sessions:
- "What approach did we take when we refactored auth last week?"\n- "Show me errors from yesterday's session"\n- "How did we handle this edge case before?"
This enables agent self-reflection - using execution history to inform\ncurrent decisions.
Built in Rust for safety and speed. 100% local, no cloud dependencies.\nThe database is just a pointer index to original logs - rebuilable anytime.
Happy to discuss the architecture or use cases.","title":"Show HN: AgTrace \u2013 Observability for AI Coding Agents via MCP (Claude Code etc.)","updated_at":"2026-03-05T23:19:59Z","url":"https://github.com/lanegrid/agtrace"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"vipdestiny"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"I've been using Claude Code, Codex CLI, and Gemini CLI on the same projects. They're each great alone but running them concurrently is chaos: merge conflicts from simultaneous edits, decisions made in one context lost to another, and slow architectural drift as each agent optimizes locally without a shared plan.
Forge is an orchestration layer that sits between AI coding tools and your codebase. It's a single Rust binary (~3 MB, zero runtime deps) that runs as an MCP server over stdio. Any MCP-compatible AI tool can call it.
What it does:\n- File locking: When an agent claims a task, Forge locks the target files. Other agents see the lock and work on something else. Conflicts become structurally impossible.\n- Knowledge flywheel: Agents call forge_capture_knowledge to store decisions, patterns, and gotchas. Other agents query the store before making decisions. Knowledge compounds across sessions instead of evaporating.\n- Drift detection: Sends recent changes + the project spec to an LLM for alignment scoring. Catches "you were supposed to build auth but you're refactoring CSS" before it compounds.\n- Governance: 5-dimension health check (tests, security, docs, architecture, git hygiene) that agents and humans can query at any time.
The brain is pluggable: a free heuristic engine (pattern matching, works offline) or an LLM engine (GPT-4.1 by default). Switch with one CLI command.
State is a single JSON file in .forge/ \u2014 human-readable, git-trackable, zero operational overhead.
51 tests (30 unit, 9 CLI, 12 MCP protocol), 0 compiler warnings, 0 unsafe blocks.
MIT licensed. Whitepaper with the full architecture: https://nxtg.ai/insights/forge-whitepaper
Happy to answer questions about the Rust implementation, MCP protocol design, or the multi-agent coordination problem in general."},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: Forge \u2013 3MB Rust binary that coordinates multi-AI coding agents via MCP"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai"],"value":"https://github.com/nxtg-ai/forge-orchestrator"}},"_tags":["story","author_vipdestiny","story_46943041","show_hn"],"author":"vipdestiny","children":[47030714],"created_at":"2026-02-09T08:38:51Z","created_at_i":1770626331,"num_comments":1,"objectID":"46943041","points":1,"story_id":46943041,"story_text":"I've been using Claude Code, Codex CLI, and Gemini CLI on the same projects. They're each great alone but running them concurrently is chaos: merge conflicts from simultaneous edits, decisions made in one context lost to another, and slow architectural drift as each agent optimizes locally without a shared plan.
Forge is an orchestration layer that sits between AI coding tools and your codebase. It's a single Rust binary (~3 MB, zero runtime deps) that runs as an MCP server over stdio. Any MCP-compatible AI tool can call it.
What it does:\n- File locking: When an agent claims a task, Forge locks the target files. Other agents see the lock and work on something else. Conflicts become structurally impossible.\n- Knowledge flywheel: Agents call forge_capture_knowledge to store decisions, patterns, and gotchas. Other agents query the store before making decisions. Knowledge compounds across sessions instead of evaporating.\n- Drift detection: Sends recent changes + the project spec to an LLM for alignment scoring. Catches "you were supposed to build auth but you're refactoring CSS" before it compounds.\n- Governance: 5-dimension health check (tests, security, docs, architecture, git hygiene) that agents and humans can query at any time.
The brain is pluggable: a free heuristic engine (pattern matching, works offline) or an LLM engine (GPT-4.1 by default). Switch with one CLI command.
State is a single JSON file in .forge/ \u2014 human-readable, git-trackable, zero operational overhead.
51 tests (30 unit, 9 CLI, 12 MCP protocol), 0 compiler warnings, 0 unsafe blocks.
MIT licensed. Whitepaper with the full architecture: https://nxtg.ai/insights/forge-whitepaper
Happy to answer questions about the Rust implementation, MCP protocol design, or the multi-agent coordination problem in general.","title":"Show HN: Forge \u2013 3MB Rust binary that coordinates multi-AI coding agents via MCP","updated_at":"2026-03-05T23:33:00Z","url":"https://github.com/nxtg-ai/forge-orchestrator"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"sys_ravi"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Sprintra \u2013 Persistent memory for AI coding agents (MCP-native)"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://sprintra.io"}},"_tags":["story","author_sys_ravi","story_47591817"],"author":"sys_ravi","children":[47591843],"created_at":"2026-03-31T18:54:02Z","created_at_i":1774983242,"num_comments":0,"objectID":"47591817","points":1,"story_id":47591817,"title":"Sprintra \u2013 Persistent memory for AI coding agents (MCP-native)","updated_at":"2026-03-31T18:56:27Z","url":"https://sprintra.io"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"cdnsteve"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Sugar gives AI coding agents a persistent, cross-project memory via MCP.
Most agents start every session knowing nothing about your codebase, your preferences, or decisions you made last week. Sugar gives them a memory that survives sessions, survives switching tools, and survives switching projects.
It stores typed memories: decisions, preferences, file context, error patterns, research notes, outcomes, and guidelines. When your agent starts a session, it can pull relevant context from ~/.sugar/memory.db - shared across all your projects. Switch from Claude Code to Goose mid-week? Same memory.
Tech: Python, SQLite + FTS5, sentence-transformers for vector search with FTS5/LIKE fallback. No server. No Docker. No cloud account.
pip install sugarai\nAGPL-3.0. GitHub: https://github.com/roboticforce/sugar
Happy to answer questions about what works, what doesn't, and where it's rough around the edges."},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: Sugar \u2013 Cross-project memory for AI coding agents via MCP"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/roboticforce/sugar/"}},"_tags":["story","author_cdnsteve","story_47415575","show_hn"],"author":"cdnsteve","children":[47538936],"created_at":"2026-03-17T17:17:31Z","created_at_i":1773767851,"num_comments":0,"objectID":"47415575","points":1,"story_id":47415575,"story_text":"Sugar gives AI coding agents a persistent, cross-project memory via MCP.
Most agents start every session knowing nothing about your codebase, your preferences, or decisions you made last week. Sugar gives them a memory that survives sessions, survives switching tools, and survives switching projects.
It stores typed memories: decisions, preferences, file context, error patterns, research notes, outcomes, and guidelines. When your agent starts a session, it can pull relevant context from ~/.sugar/memory.db - shared across all your projects. Switch from Claude Code to Goose mid-week? Same memory.
Tech: Python, SQLite + FTS5, sentence-transformers for vector search with FTS5/LIKE fallback. No server. No Docker. No cloud account.
pip install sugarai\nAGPL-3.0. GitHub: https://github.com/roboticforce/sugar
Happy to answer questions about what works, what doesn't, and where it's rough around the edges.","title":"Show HN: Sugar \u2013 Cross-project memory for AI coding agents via MCP","updated_at":"2026-03-27T19:41:55Z","url":"https://github.com/roboticforce/sugar/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"moonx010"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"I built an MCP server that gives AI coding agents persistent memory across projects.
The problem: tools like Claude Code and Codex have per-project memory, butwhen you switch between projects, the agent starts from scratch. "How was that API project going?" \u2014 it has no idea.
Hive Memory solves this by acting as a central memory layer. The agent automatically stores decisions, learnings, and session progress, then restores context when you come back.
Key points:\n- 14 MCP tools: project management, memory store/recall, session save, group organization\n- Works with Claude Code, Claude Desktop, Cursor, Codex \u2014 any MCP client\n- Auto-discovers projects in a directory and onboards them\n- Group system for organizing related projects with shared guides\n- Fully local (stored in ~/.cortex/ as JSON/Markdown). No cloud, no accounts\n- npm install -g hive-memory
The typical workflow: open a project \u2192 agent loads last session context \u2192 work \u2192 agent saves progress \u2192 switch to another project \u2192 repeat. Days later, come back and the agent knows exactly where you left off.
Built with TypeScript + MCP SDK. Optional Rust native module for semantic search."},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: Hive Memory \u2013 Cross-project memory for AI coding agents (MCP server)"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/moonx010/hive-memory"}},"_tags":["story","author_moonx010","story_47207442","show_hn"],"author":"moonx010","children":[47215624],"created_at":"2026-03-01T15:15:36Z","created_at_i":1772378136,"num_comments":0,"objectID":"47207442","points":1,"story_id":47207442,"story_text":"I built an MCP server that gives AI coding agents persistent memory across projects.
The problem: tools like Claude Code and Codex have per-project memory, butwhen you switch between projects, the agent starts from scratch. "How was that API project going?" \u2014 it has no idea.
Hive Memory solves this by acting as a central memory layer. The agent automatically stores decisions, learnings, and session progress, then restores context when you come back.
Key points:\n- 14 MCP tools: project management, memory store/recall, session save, group organization\n- Works with Claude Code, Claude Desktop, Cursor, Codex \u2014 any MCP client\n- Auto-discovers projects in a directory and onboards them\n- Group system for organizing related projects with shared guides\n- Fully local (stored in ~/.cortex/ as JSON/Markdown). No cloud, no accounts\n- npm install -g hive-memory
The typical workflow: open a project \u2192 agent loads last session context \u2192 work \u2192 agent saves progress \u2192 switch to another project \u2192 repeat. Days later, come back and the agent knows exactly where you left off.
Built with TypeScript + MCP SDK. Optional Rust native module for semantic search.","title":"Show HN: Hive Memory \u2013 Cross-project memory for AI coding agents (MCP server)","updated_at":"2026-03-05T23:39:02Z","url":"https://github.com/moonx010/hive-memory"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gajus"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"I started to use AI agents for coding and quickly ran into a frustrating limitation \u2013 there is no easy way to share my development environment logs with AI agents. So that's what is Teemux. A simple CLI program that aggregates logs, makes them available to you as a developer (in a pretty UI), and makes them available to your AI coding agents using MCP.
There is one implementation detail that I geek out about:
It is zero config and has built-in leader nomination for running the web server and MCP server. When you start one `teemux` instance, it starts web server, .. when you start second and third instances, they join the first server and start merging logs. If you were to kill the first instance, a new leader is nominated. This design allows to seamless add/remove nodes that share logs (a process that historically would have taken a central log aggregator).
A super quick demo:
npx teemux -- curl -N https://teemux.com/random-logs"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["mcp"],"value":"Show HN: Teemux \u2013 Zero-config log multiplexer with built-in MCP server"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://teemux.com/"}},"_tags":["story","author_gajus","story_46733905","show_hn"],"author":"gajus","children":[46734213,46736688,46737065,46743460],"created_at":"2026-01-23T15:49:35Z","created_at_i":1769183375,"num_comments":9,"objectID":"46733905","points":11,"story_id":46733905,"story_text":"I started to use AI agents for coding and quickly ran into a frustrating limitation \u2013 there is no easy way to share my development environment logs with AI agents. So that's what is Teemux. A simple CLI program that aggregates logs, makes them available to you as a developer (in a pretty UI), and makes them available to your AI coding agents using MCP.
There is one implementation detail that I geek out about:
It is zero config and has built-in leader nomination for running the web server and MCP server. When you start one `teemux` instance, it starts web server, .. when you start second and third instances, they join the first server and start merging logs. If you were to kill the first instance, a new leader is nominated. This design allows to seamless add/remove nodes that share logs (a process that historically would have taken a central log aggregator).
A super quick demo:
npx teemux -- curl -N https://teemux.com/random-logs","title":"Show HN: Teemux \u2013 Zero-config log multiplexer with built-in MCP server","updated_at":"2026-03-05T23:24:06Z","url":"https://teemux.com/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"vipdestiny"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Forge \u2013 A 3 MB Rust binary that coordinates multiple AI coding agents via MCP\nhttps://github.com/nxtg-ai/forge-orchestrator
Forge is an orchestration layer that sits between AI coding tools and your codebase. It's a single Rust binary (~3 MB, zero runtime deps) that runs as an MCP server over stdio. Any MCP-compatible AI tool can call it.
MIT licensed. Whitepaper with the full architecture: https://nxtg.ai/insights/forge-whitepaper"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What are you working on? (February 2026)"}},"_tags":["comment","author_vipdestiny","story_46937696"],"author":"vipdestiny","comment_text":"Forge \u2013 A 3 MB Rust binary that coordinates multiple AI coding agents via MCP\nhttps://github.com/nxtg-ai/forge-orchestrator
Forge is an orchestration layer that sits between AI coding tools and your codebase. It's a single Rust binary (~3 MB, zero runtime deps) that runs as an MCP server over stdio. Any MCP-compatible AI tool can call it.
MIT licensed. Whitepaper with the full architecture: https://nxtg.ai/insights/forge-whitepaper","created_at":"2026-02-09T08:34:47Z","created_at_i":1770626087,"objectID":"46943009","parent_id":46937696,"story_id":46937696,"story_title":"Ask HN: What are you working on? (February 2026)","updated_at":"2026-03-05T23:33:00Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"zannic"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Trust Boundary Failures in AI Coding Agents (Claude Code MCP Attacks)"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://zenodo.org/records/19011781"}},"_tags":["story","author_zannic","story_47372349"],"author":"zannic","created_at":"2026-03-14T01:29:03Z","created_at_i":1773451743,"num_comments":0,"objectID":"47372349","points":1,"story_id":47372349,"title":"Trust Boundary Failures in AI Coding Agents (Claude Code MCP Attacks)","updated_at":"2026-03-14T19:33:30Z","url":"https://zenodo.org/records/19011781"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"oleksiijko"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"PMB is a local-first memory system for AI coding agents, running on MCP and storing data with SQLite and LanceDB. It preserves project context, decisions, and preferences across sessions and tools. I created it because coding agents frequently lose valuable context between sessions, especially on long projects. So, I'd love feedback from those who use coding agents regularly. What do you think should be auto-remembered versus what should always need explicit confirmation?"},"title":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["ai","coding","agent"],"value":"Show HN: PMB \u2013 Local-first memory for AI coding agents"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/oleksiijko/pmb"}},"_tags":["story","author_oleksiijko","story_48415595","show_hn"],"author":"oleksiijko","created_at":"2026-06-05T17:24:31Z","created_at_i":1780680271,"num_comments":0,"objectID":"48415595","points":2,"story_id":48415595,"story_text":"PMB is a local-first memory system for AI coding agents, running on MCP and storing data with SQLite and LanceDB. It preserves project context, decisions, and preferences across sessions and tools. I created it because coding agents frequently lose valuable context between sessions, especially on long projects. So, I'd love feedback from those who use coding agents regularly. What do you think should be auto-remembered versus what should always need explicit confirmation?","title":"Show HN: PMB \u2013 Local-first memory for AI coding agents","updated_at":"2026-06-07T03:30:55Z","url":"https://github.com/oleksiijko/pmb"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"crestfallen33"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Hey HN! My name is Nick, and I lead Product Marketing at Cline. Over the past few weeks, we\u2019ve been hard at work on a major update for Cline (https://github.com/cline/cline) and would love your feedback.
For those who haven\u2019t heard of us, Cline is an open-source AI coding agent extension for VS Code (we just hit 30k stars on GitHub). It\u2019s model agnostic and works on a bring-your-own-key API model, enabling Cline to read, write, and edit your files so you can build and maintain complex software faster than ever before.
A couple of months back, we introduced MCP servers (https://modelcontextprotocol.io/introduction) to Cline, allowing users to integrate new tools powered by Anthropic\u2019s Model Context Protocol. While we\u2019ve seen some really cool uses\u2014from perplexity research to Notion integration\u2014we noticed that many users find the concept of \u201cMCP\u201d a bit abstract.
That\u2019s why we just launched the MCP Marketplace\u2014a one-click install suite where you can easily add a variety of new tools to Cline. We believe this will lower the barrier to entry and help you better understand and leverage MCP servers. (Tip: The perplexity MCP is my favorite!)
I\u2019d love to hear from both long-time Cline users and newcomers. Your feedback will help shape our content and support efforts for adopting MCP servers in Cline.
Here are a few questions that would really help us improve Cline:
1. What is your current understanding of MCP servers? \n - How do you think they work and what benefits do you expect?
2. Have you tried using MCP servers in Cline?\n - If yes, what was your experience like? If not, what\u2019s holding you back?
3. Where do you see gaps in understanding or setup with MCP servers?\n - Are there particular aspects (documentation, installation, use cases) that feel unclear or challenging?
4. What kind of content or support would help you adopt MCP servers more confidently? \n - For example, would detailed tutorials, videos, or troubleshooting guides be useful?
Looking forward to your thoughts and questions\u2014thanks for taking the time to help us build a better tool! Thank you!
-Nick & the Cline team"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: Cline \u2013 Open-Source VS Code AI Coding Agent with a New MCP Marketplace"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/cline/cline"}},"_tags":["story","author_crestfallen33","story_43105538","show_hn"],"author":"crestfallen33","children":[43116346],"created_at":"2025-02-19T18:29:07Z","created_at_i":1739989747,"num_comments":1,"objectID":"43105538","points":3,"story_id":43105538,"story_text":"Hey HN! My name is Nick, and I lead Product Marketing at Cline. Over the past few weeks, we\u2019ve been hard at work on a major update for Cline (https://github.com/cline/cline) and would love your feedback.
For those who haven\u2019t heard of us, Cline is an open-source AI coding agent extension for VS Code (we just hit 30k stars on GitHub). It\u2019s model agnostic and works on a bring-your-own-key API model, enabling Cline to read, write, and edit your files so you can build and maintain complex software faster than ever before.
A couple of months back, we introduced MCP servers (https://modelcontextprotocol.io/introduction) to Cline, allowing users to integrate new tools powered by Anthropic\u2019s Model Context Protocol. While we\u2019ve seen some really cool uses\u2014from perplexity research to Notion integration\u2014we noticed that many users find the concept of \u201cMCP\u201d a bit abstract.
That\u2019s why we just launched the MCP Marketplace\u2014a one-click install suite where you can easily add a variety of new tools to Cline. We believe this will lower the barrier to entry and help you better understand and leverage MCP servers. (Tip: The perplexity MCP is my favorite!)
I\u2019d love to hear from both long-time Cline users and newcomers. Your feedback will help shape our content and support efforts for adopting MCP servers in Cline.
Here are a few questions that would really help us improve Cline:
1. What is your current understanding of MCP servers? \n - How do you think they work and what benefits do you expect?
2. Have you tried using MCP servers in Cline?\n - If yes, what was your experience like? If not, what\u2019s holding you back?
3. Where do you see gaps in understanding or setup with MCP servers?\n - Are there particular aspects (documentation, installation, use cases) that feel unclear or challenging?
4. What kind of content or support would help you adopt MCP servers more confidently? \n - For example, would detailed tutorials, videos, or troubleshooting guides be useful?
Looking forward to your thoughts and questions\u2014thanks for taking the time to help us build a better tool! Thank you!
-Nick & the Cline team","title":"Show HN: Cline \u2013 Open-Source VS Code AI Coding Agent with a New MCP Marketplace","updated_at":"2025-05-12T19:46:47Z","url":"https://github.com/cline/cline"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"hackeruse"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"WireMock MCP: Use AI Coding Agents for Codebase-Aware API Simulation"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"https://www.wiremock.io/post/wiremock-mcp-ai-coding-agents"}},"_tags":["story","author_hackeruse","story_43557219"],"author":"hackeruse","children":[43557233],"created_at":"2025-04-02T14:39:01Z","created_at_i":1743604741,"num_comments":0,"objectID":"43557219","points":2,"story_id":43557219,"title":"WireMock MCP: Use AI Coding Agents for Codebase-Aware API Simulation","updated_at":"2025-04-02T14:58:56Z","url":"https://www.wiremock.io/post/wiremock-mcp-ai-coding-agents"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"JulianQuinn"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"Show HN: TaskPeace \u2013 a task queue my AI coding agents pull work from over MCP"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://taskpeace.com/"}},"_tags":["story","author_JulianQuinn","story_48775484","show_hn"],"author":"JulianQuinn","children":[48775538,48776775,48777114,48777532,48784080,48794997],"created_at":"2026-07-03T14:27:26Z","created_at_i":1783088846,"num_comments":7,"objectID":"48775484","points":7,"story_id":48775484,"title":"Show HN: TaskPeace \u2013 a task queue my AI coding agents pull work from over MCP","updated_at":"2026-07-08T21:56:12Z","url":"https://taskpeace.com/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ibrarmaikah"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"WSP WordPress MCP \u2013 Connect AI Coding Agents to WordPress"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["mcp"],"value":"https://github.com/bilalnaseer/wsp-wordpress-mcp"}},"_tags":["story","author_ibrarmaikah","story_48474367"],"author":"ibrarmaikah","created_at":"2026-06-10T10:48:12Z","created_at_i":1781088492,"num_comments":0,"objectID":"48474367","points":4,"story_id":48474367,"title":"WSP WordPress MCP \u2013 Connect AI Coding Agents to WordPress","updated_at":"2026-06-11T11:51:13Z","url":"https://github.com/bilalnaseer/wsp-wordpress-mcp"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ibrarmaikah"},"title":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["ai","coding","agent","mcp"],"value":"WSP WordPress MCP \u2013 Connect AI Coding Agents to WordPress"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["mcp"],"value":"https://github.com/bilalnaseer/wsp-wordpress-mcp"}},"_tags":["story","author_ibrarmaikah","story_48488856"],"author":"ibrarmaikah","children":[48525032],"created_at":"2026-06-11T11:20:43Z","created_at_i":1781176843,"num_comments":1,"objectID":"48488856","points":2,"story_id":48488856,"title":"WSP WordPress MCP \u2013 Connect AI Coding Agents to WordPress","updated_at":"2026-06-14T07:36:55Z","url":"https://github.com/bilalnaseer/wsp-wordpress-mcp"}],"hitsPerPage":20,"nbHits":851,"nbPages":43,"page":0,"params":"query=AI+coding+agent+MCP&advancedSyntax=true&analyticsTags=backend","processingTimeMS":25,"processingTimingsMS":{"_request":{"queue":13,"roundTrip":14},"afterFetch":{"format":{"highlighting":1,"total":1},"merge":{"mergeLoop":{"prepareNextHit":1,"total":1},"total":1},"total":1},"fetch":{"query":5,"scanning":17,"total":23},"total":25},"query":"AI coding agent MCP","serverTimeMS":39}