Releases·Tool Release
Show HN: Plurnk, an open-source AI coding harness with model-curated context and 200MB footprint
A developer released Plurnk, an open-source AI harness dogfooded for weeks as a daily driver on an RTX 5070 Ti with 16GB VRAM. It emphasizes model-managed context curation, an ANTLR grammar for its verbs, and universal resource addressing via pseudo-URIs. The author seeks feedback and targets builders running agentic workflows under local, modest hardware constraints.
1 engagement1 sourceTue, Sep 15, 2026, 11:57 PM
AI Overview
- Plurnk's model deterministically curates its own context via a dozen markdown-like verbs defined in an ANTLR EBNF grammar
- Treesitter-built repo graph searchable via glob, regex, path, xpath, graph, and SQLite fulltext; everything addressable through pseudo-URIs
- Supports MCP2, AG-UI, A2A, and OAI spec; offers CLI, TUI, Neovim plugin, and a 200MB service/client build for weak hardware
- Developer dogfoods it as a daily driver on an RTX 5070 Ti with 16GB VRAM; sysprompt kept near pi agent's size
Entities
Related
Social·Community
User seeks agentic harness for small local models
A semi-beginner local AI user asks for recommendations on an agentic harness that fits small models and modest hardware, listing requirements like web search, browser use, MCP servers, and a sandboxed code environment. The post highlights a gap in tooling for lightweight local agent setups.
A semi-beginner local AI user asks for recommendations on an agentic harness that fits small models and modest hardware, listing requirements like web search, browser use, MCP servers, and a sandboxed code environment. The post highlights a gap in tooling for lightweight local agent setups.
83 engagement1 sourceredditSat, Aug 15, 2026, 07:59 PM
Products·building
Crucible – terminal AI agent harness powered by custom functional logic language Obverse
Crucible is a terminal-native AI agent harness driven by Obverse, a custom Verse/Oz-inspired functional logic programming language with a built-in constraint solver. It lets LLM agents offload deterministic constraint solving and state management instead of hand-rolling loops, regex, and state-machine wrappers. It targets developers building long-running agent sessions where async boilerplate breaks down over time.
Crucible is a terminal-native AI agent harness driven by Obverse, a custom Verse/Oz-inspired functional logic programming language with a built-in constraint solver. It lets LLM agents offload deterministic constraint solving and state management instead of hand-rolling loops, regex, and state-machine wrappers. It targets developers building long-running agent sessions where async boilerplate breaks down over time.
1 engagement1 sourceredditFri, Sep 4, 2026, 08:22 PM
Releases·Tool Release
Dev releases Brain, an open-source minimal agent runtime pitched against LangChain and Mastra (<1ms session, ~25MB binary)
A developer open-sourced 'brain' (github.com/aexhq/brain), a minimal, extensible agent runtime inspired by Anthropic's managed agents engineering post and pi.dev. It targets production agent use cases the author found too rigid in LangChain and Mastra. For practitioners, it offers a lightweight alternative with sub-second cold starts and a small compiled footprint.
A developer open-sourced 'brain' (github.com/aexhq/brain), a minimal, extensible agent runtime inspired by Anthropic's managed agents engineering post and pi.dev. It targets production agent use cases the author found too rigid in LangChain and Mastra. For practitioners, it offers a lightweight alternative with sub-second cold starts and a small compiled footprint.
2 engagement1 sourceredditMon, Sep 7, 2026, 10:08 PM
Releases·Tool Release
Dev builds Rust REPL/DSL and tiny MoE models running in-browser and on ARM/RISC-V boards to visualize MoE experts and Engram
An ML study-group member shared a browser-based 'MoE Microscope' built with a custom Rust REPL/array-programming DSL, demonstrating tiny dense and MoE models to visualize experts, routing, and Engram-style concepts. The tool targets browser (WASM), macOS/Linux CLI with Apple and NVIDIA GPUs, and eventually small ARM/RISC-V boards.
An ML study-group member shared a browser-based 'MoE Microscope' built with a custom Rust REPL/array-programming DSL, demonstrating tiny dense and MoE models to visualize experts, routing, and Engram-style concepts. The tool targets browser (WASM), macOS/Linux CLI with Apple and NVIDIA GPUs, and eventually small ARM/RISC-V boards.
2 engagement1 sourcehackernewsMon, Sep 14, 2026, 01:59 AM
Releases·Tool Release
awesome-ai-harness: curated list for harness engineering
A new GitHub repository, awesome-ai-harness, curates resources on harness engineering—the runtime scaffolding that turns language models into working agents. It covers context management, tool design, agent loops, memory, sandboxing, and evals, with an English and Chinese version.
A new GitHub repository, awesome-ai-harness, curates resources on harness engineering—the runtime scaffolding that turns language models into working agents. It covers context management, tool design, agent loops, memory, sandboxing, and evals, with an English and Chinese version.
50 engagement1 sourcegithubSun, Aug 16, 2026, 04:57 AM