Binh Tran

HTML vs Markdown for agent memory. Which is better in real production? After the previous benchmark with 603 questions and HTML won on all accuracy, token saving and latency reduction. So, we were questioning: Does HTML still perform better in real production at a larger scale? To answer that, we re-ran our LoCoMo benchmark on the full set with 1,982 questions across 11 conversations. The results: - HTML scored 90.77% accuracy at a total cost of $4.11, while Markdown scored 90.51% at $8.30. - HTML ran 40% faster on query and 12.5% faster on curate. Beyond the numbers, HTML brings three more benefits: 1. Anyone with a browser can open HTML without extra tools: An HTML file opens directly in any browser, while Markdown files need a separate viewer or editor to render properly. 2. Agents can search HTML like a database and return the exact answers 3. HTML fits inside the new cross-agent protocols: A2A (Agent-to-Agent) and MCP (Model Context Protocol) both expect standardized input and output, and HTML provides exactly that. We are shipping HTML as the default format for ByteRover soon

HTML vs Markdown for agent memory: Which is cheaper? Before we ran the benchmark, we expected HTML to cost more because it carries images and charts. Surprisingly it was 42% cheaper. Also 5.9% more accurate and 39% faster The test was conducted across 271 sessions, 603 questions on LoCoMo Check how we tested ↓

Your memory, organizing itself while you sleep. ByteRover Dreaming runs while your agent is idle: → merges near-duplicate notes → connects topics that share patterns → archives stale drafts Nothing changes without you. Every proposal comes with the reasoning, and one command to approve. run brv dream

Your shared memory across Claude Code, OpenClaw, and Hermes is invisible. Now it's a webpage. The Local Web UI for ByteRover (OSS memory system): Manage your team's context across every entry and every project, in one place. → trace decisions back to where they were made → recall how a teammate solved the same problem last month → see context ranked by relevance, automatically

Never let one vendor own your workflow" is trending again today. the version that actually matters: never let one vendor own your context. models are swappable. chat history, prompt libraries, custom instructions, integration state, agent memory those you can't rebuild when the vendor goes sideways. what you accumulated is the moat. keeping it portable is the discipline. the vendor risk everyone's scared of isn't losing the model. it's losing everything you built on top of it

Build Your Own Autonomous Local Agent Stacks That save 83% on token costs and give your agents 92% long-term memory retention. This guide helps you set up a fully local and private autonomous stack. This means your data stays on your machine, your agents don't forget their tasks, and you stop paying for expensive cloud API tokens. The Stack Components: 💪The Execution Agents: OpenClaw (2026.4.12)/ Hermes → run the actual tasks and scripts on your computer. 🧠The Brain (Local LLM): Gemma 4, Z.ai’s GLM-5.1, and Qwen 3.5 → able to handle complex tasks like coding, media generation, or researching 🤖The Memory Engine: ByteRover (3.3.0) → filesystem memory that connects natively with your tools and keeps context organized Hardware Requirements - For this experimental setup, we run on Mac M4 Pro, RAM 24GB (ByteRover and autonomous agents can run on a Apple RAM 24GB machine) - For production usage we recommend at least an Apple M4 with RAM 48GB. Step 1: Downloads the local LLM models You will need two specific models to balance performance and memory efficiency: - Gemma 4 E4B (Q4): This is for OpenClaw to handle reasoning. (Uses ~8.7 GB RAM) - Qwen 3.5-9B (Q8): This is for ByteRover to manage your data. (Uses ~10.5 GB RAM) On a 24 GB machine, both models fit in memory simultaneously. Gemma 4 E4B at Q4 uses ~8.7 GB and Qwen3.5-9B at Q8 uses ~10.5 GB. Step 2: Load Both Models in LM Studio Open LM Studio and load both models simultaneously. On a 24GB machine, these will fit together, leaving just enough room for your system to run smoothly. Step 3: Configure Your Agent Whether you choose OpenClaw or Hermes, the setup is the same: - Point the agent to your local LM Studio endpoint. - Adjust Memory: By default, OpenClaw handles 50,000 tokens. To change this, edit your openclaw.json file and run openclaw gateway restart to apply the update. Step 4: Connect ByteRover Memory Finally, set up the ByteRover CLI: - Link it to the same local endpoint. - Select the Qwen model as the primary memory handler. The Result: You now have a persistent AI assistant that remembers your project details across different sessions. Everything stays local, with Cloud option if you ever need to collab with your teams Detailed guide for each step 👇

@ashwingop tested five architecturally distinct memory systems (vector DBs, knowledge graphs, context windows, BM25/filesystem, parametric memory) and then proved with math why the filesystem approach wins. He mentions @ByteroverDev (alongside with @Letta_AI , ClaudeCode, @ManusAI and @openclaw) as rare real-world examples of filesystem-based memory actually shipping in production and delivering meaningful accuracy. Here’s what we’ve been shipping: - Your memory lives in simple, human-readable markdown files (not hidden vectors) - Native memory for OpenClaw, Hermes, and 22+ AI coding agents including ClaudeCode and Cursor. Plus, native Obsidian support - More than 92% on LoCoMo & LongMemEval benchmark - Portable by default (local-first), with optional cloud sync for team collaboration and enterprise-grade security Huge respect for Ashwin Gopinath for formalizing into a full paper arxiv.org/html/2603.27116v1. This is the kind of rigorous, formal analysis that moves the entire agent memory field forward. If you’re building agents that need to actually remember long-term, check out this repo github.com/campfirein/bytero… Install the CLI: curl -fsSL <byterover.dev/install.sh> | sh

Introduce Context Intelligence - Your LLM Knowledge Wiki Now Knows What Knowledge Matters to You (with Smart Importance Score) Imagine your agents like OpenClaw, Hermes, ClaudeCode have intelligence score to know what knowledge matters We just shipped Context Intelligence - the upgrade that turns your knowledge base from a simple storage folder into a living, learning system that actually understands what matters to you. We added three new intelligence layers across the context lifecycle: - Compression: retrieved knowledge now comes with compact summaries, so both agents and humans can scan first and expand only when needed. - Preservation: important decisions and session learnings are extracted into structured memory, including decisions, patterns, entities, skills, and preferences. - Prioritization: knowledge now gets ranked by importance, helping future retrieval surface what matters most instead of treating every note equally. Why this matters: In practice, workflows don't usually fail because raw context is too dense and disorganized → High token costs, agents return incorrect information These new layers bridge that gap. We’ve integrated: - Summaries for readability - Structured extraction for durable memory - Decision timelines to preserve intent - Importance scoring to prioritize and surface the exact knowledge

Manage AI Agent Knowledge like Git (with Version Control) When your whole team uses and codes across agents like OpenClaw, Hermes, Claude Code -> Manage shared knowledge & Keep that knowledge accurate and up-to-date gets surprisingly hard. Why? Because code evolves and teams grow. Different people and different agent sessions interpret the same codebase in slightly different ways. Until now with ByteRover, syncing the context tree was just a simple “upload and download” to the cloud. There’s no history, no way to see what changed or why. The last person (or agent) to push simply overwrote everything. Today we fixed that to make Memory/ Context Management Fully Git-like with Version Control. We just released 15 Git-like operations for your shared context tree. This turns your AI’s memory from a messy shared folder into a proper, version-controlled knowledge base. The biggest value? You can now compose memory exactly the way you want, deciding precisely which insights, rules, and learnings become shared knowledge for the whole team, and which stay private until they’re ready. Here’s how it works:

ByteRover just shipped Context Management - the feature that allow you to browse, organize, and search your project's context data. The Idea: We take Karpathy’s LLM Knowledge Base wikis (Obsidian + LLM workflow) to the next level: A product that’s free, open-sourced, production-proven and you can share with your entire team’s code, docs, and context: - LLM automatically stores, organizes all knowledge/ context for you, recalls when you need. It cleans, tags, summarizes, removes outdated knowledge and keeps your whole team’s knowledge base tidy without any manual work. - You can easily view & manage your whole team’s context. Raw context → organized, beautiful, filterable, searchable, and versioned. - Your project context will become a smart, queryable that grows smarter every time you use it. This new feature Context Management allows you to easily view your context It introduces two visual modes: Tree View and Grid View along with a powerful full-text search with filtering capabilities.