Search Results for "shared memory allocator"
Sort By:
Showing 33 open source projects for "shared memory allocator"
View related business solutions-
Paessler: Easy to Use With Enterprise Power. Free TrialYou shouldn't have to choose between a monitoring tool that's easy to use and one that's powerful enough for a complex environment. PRTG's low-code interface lets any admin build dashboards, set alerts and monitor devices without scripting, while custom sensors and full API access are there when your team needs deeper control. One platform, no compromise. Download a free 30-day trial now.
-
Train ML Models With SQL You Already KnowBuild and deploy ML models using familiar SQL. Automate data prep with built-in Gemini. Query 1 TB and store 10 GB free monthly.
-
1
codebase-memory-mcp
High-performance code intelligence MCP server
...It is built for speed, with average repositories indexed in milliseconds and structural queries answered in under one millisecond. It ships as a single static binary for macOS, Linux, and Windows with no shared library dependencies. It also includes semantic search, BM25 search, graph visualization, auto-indexing, and cross-repository intelligence. -
2
OpenSquilla
Token-Efficient AI Agent with same budget, higher intelligence density
OpenSquilla is a token-efficient microkernel AI agent runtime designed for CLI, web UI, and chat-based workflows. It routes each turn through a shared loop that can select lower-cost models when appropriate while preserving tool dispatch, retries, memory, and decision logging. The project supports multiple LLM providers through a pluggable provider layer, making it adaptable to different model ecosystems. It includes persistent memory, built-in web search, on-device embeddings, and sandboxing for safer execution. ... -
3
Cumora
Where agent teams gather. Cross-platform team chat
Cumora is a cross-platform team chat where humans and AI agents participate as peers in conversations and shared work. Agents can maintain personas and memory, claim tasks, coordinate with one another, send and receive email, and use shared Kanban and calendar views. Teams can run agents through Cumora Cloud or bring their own Claude Code or Codex environment from a Mac or VPS. Cloud agents receive isolated runtime pods with access to tools such as shell commands, files, browsers, email, memory, and skills. ... -
4
Hivemind
Hivemind turns your traces into reusable skills across agents
Hivemind is a shared memory and skill system for AI coding agents. It captures agent sessions as structured traces, including prompts, tool calls, responses, and subagent activity. The system mines those traces for repeated patterns and turns them into reusable skills that can be shared across a team. It supports agents such as Claude Code, OpenClaw, Codex, Cursor, Hermes, and pi. -
Custom VMs From 1 to 96 vCPUs With 99.95% UptimeLive migration and automatic failover keep workloads online through maintenance. One free e2-micro VM every month.
-
5
EverOS
One portable memory layer for every AI agent
...The system is designed to preserve user context and agent experience across sessions without forcing dependence on a hosted service. EverOS can serve as shared memory infrastructure for coding assistants, personal agents, and custom applications. Its architecture emphasizes user ownership, portability, and self-evolving reuse of accumulated knowledge. -
6
Rakazo
Open-source Grok Bot alternative. Choose your own model and sandbox
Rakazo is an open-source platform for running persistent AI teammates across web, desktop, and mobile clients. Each bot maintains its own conversations, memory, routines, and history instead of acting as a disposable chat session. Bots can use browsers, terminals, files, and graphical desktops through shared team computers or isolated private environments. They can delegate work to peer bots or temporary subagents and can use user-provided model credentials. Integrations are available through Composio, Pipedream Connect, remote MCP servers, OpenAPI tools, and other supported sources. ... -
7
Alook
The collaboration layer for your AI workforce
...Alook also emphasizes memory and traceability, so agents can remember past decisions, learn preferences, and keep a record of instructions and replies. Its main value is turning separate AI coding tools into an organized, always-on operating system for agent-based work. -
8
QM
Multiplayer agent harness for work
QM is an open-source multiplayer agent harness designed for startup teams and company-wide AI work. Each person, channel, and project receives its own scoped memory, files, credentials, permissions, schedules, web apps, and durable sandbox. The same agent identity and configuration can operate through Slack or the web interface. Teams can choose among supported models and harnesses, including Pi, OpenCode, Codex, and Claude Code, without binding the deployment to one vendor. Administrators control organization settings, available models, shared skills, and security posture. ... -
9
Strata
Qwen3.8-Flash-Next on any consumer hardware
...Optional image input and several quantized model sizes allow users to balance memory use, speed, and quality. -
Host LLMs in Production With On-Demand GPUsDeploy your model, get an endpoint, pay only for compute time. No GPU provisioning or infrastructure management required.
-
10
Munder Difflin
local multi-agent harness
Munder Difflin is a desktop multi-agent harness that turns terminal coding CLIs into a coordinated team of autonomous workers. Each supported CLI runs as a real terminal process with its own identity, workspace, and visual avatar. A central orchestrator called Michael routes tasks, resolves routine coordination, and escalates sensitive actions for human approval. Agents communicate through local mailboxes, share a blackboard, and retain long-term memory across sessions. The app provides live... -
11
TurboFieldfare
Gemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook
TurboFieldfare is a custom Swift and Metal runtime for running the instruction-tuned Gemma 4 26B-A4B model on Apple Silicon Macs with limited memory. Instead of loading the entire model, it keeps the shared core and KV cache in RAM while streaming only the routed experts required for each token from SSD. This approach reduces active memory use to roughly 2 GB while the installed model occupies about 14.3 GB of storage. The project includes a native Mac application, command-line tools, a streaming installer, a Swift library, and an experimental OpenAI-compatible local server. ... -
12
METATRON
AI-powered penetration testing assistant using local LLM on linux
METATRON is a multi-agent AI orchestration framework designed to coordinate complex workflows across multiple intelligent agents. It provides a structured system for task delegation, communication, and collaboration between agents. The framework emphasizes scalability, allowing multiple agents to work together on large or complex problems. It includes mechanisms for managing context, memory, and execution flow across tasks. METATRON is particularly useful for building advanced AI systems... -
13
WeChat-AI
Self-hosted WeChat role-playing conversation service
WeChat-AI is a self-hosted roleplay chatbot service built for WeChat and connected directly through Tencent iLink. Users authenticate through LINUX DO OAuth and can add bots by scanning QR codes. Each bot can receive a persona, memory, custom model settings, and access to web search through a separate tools gateway. Approved users can also establish peer-to-peer conversations that relay messages through WeChat. Administrators receive dashboards for users, bots, token usage, deployment nodes, and sticker moderation. The platform supports visual Chatflow orchestration, multi-node deployment with shared Redis, Cloudflare load balancing, and incremental OTA updates. -
14
FramePack
Lets make video diffusion practical
FramePack explores compact representations for sequences of image frames, targeting tasks where many near-duplicate frames carry redundant information. The idea is to “pack” frames by detecting shared structure and storing differences efficiently, which can accelerate training or inference on video-like data. By reducing I/O and memory bandwidth, datasets become lighter to load while models still see the essential temporal variation. The repository demonstrates both packing and unpacking steps, making it straightforward to integrate into preprocessing pipelines. ... -
15
CogVideo
Text and image to video generation: CogVideoX and CogVideo
CogVideo is an open-source family of advanced video generation models that can create videos from text, images, or existing video inputs. Built on large-scale Transformer and diffusion architectures, it enables multimodal generation across text-to-video, image-to-video, and video continuation tasks. The latest CogVideoX models offer higher resolution outputs, longer video durations, and improved controllability through prompt engineering. The project includes tools for inference,... -
16
Context Hub
Makes coding agents get smarter with every task
Context Hub is a curated documentation system built to help coding agents write more accurate code. It gives agents versioned, language-specific reference material instead of forcing them to rely on noisy web searches or stale model memory. The project includes a CLI called chub that agents can use to search for available docs, fetch specific API guidance, and request only the files they need. It also supports local annotations, allowing an agent to remember project-specific notes, pitfalls, or workarounds across future sessions. Feedback can be sent back to maintainers so shared documentation improves over time. ... -
17
Qwen
The official repo of Qwen chat & pretrained large language model
Qwen is a series of large language models developed by Alibaba Cloud, consisting of various pretrained versions like Qwen-1.8B, Qwen-7B, Qwen-14B, and Qwen-72B. These models, which range from smaller to larger configurations, are designed for a wide range of natural language processing tasks. They are openly available for research and commercial use, with Qwen's code and model weights shared on GitHub. Qwen's capabilities include text generation, comprehension, and conversation, making it a... -
18
WASTE
Run the full 2.78-trillion-parameter Kimi K3 model
WASTE is an embeddable C inference engine for running extremely large mixture-of-experts models when the weights exceed available RAM. It keeps the shared model trunk in memory and streams only the experts selected for each token from fast NVMe storage. A bounded cache reuses recently needed experts, while lookahead routing begins disk reads before the next layer requires them. Its main target is the full 2.78-trillion-parameter Kimi K3 model, including multimodal image input. The engine has no third-party runtime dependency on its CPU inference path and exposes both a CLI and C library. ... -
19
BigMac
An open-source toolkit for BigMac-style pipeline-parallel training
BigMac is an open-source toolkit for pipeline-parallel training of multimodal large language models. It preserves optimized language-model pipeline schedules while placing encoder and generator work around them. This design reduces activation memory without bringing back cross-module pipeline bubbles. Its scheduler creates global operator plans, while its executor runs those plans through a shared schedule abstraction. A Megatron-Core reference backend and Qwen3 and Qwen3-VL tutorials help developers connect the approach to real training workflows. The simulator lets researchers visualize schedules, compare pipeline strategies, and model timing imbalances caused by compute cost, input size, or uneven stage partitioning. ... -
20
Oh My codeX (OMX)
Your codex is not alone. Add hooks, agent teams, HUDs
Oh My codeX (OMX) is a multi-agent orchestration layer designed to extend the capabilities of OpenAI Codex CLI by introducing structured teamwork, automation, and advanced workflow management. It addresses limitations in the base Codex environment, such as the lack of hooks, agent coordination, and persistent execution, by layering a shell-based system that enables richer interaction patterns. The project transforms a single AI coding assistant into a coordinated system of specialized agents... -
21
Extractous
Fast and efficient unstructured data extraction
...Its purpose is to extract text and metadata efficiently from formats such as PDF, Word, HTML, email archives, images, and more, without depending on external APIs or separate parsing servers. The project emphasizes performance and low memory usage, and its maintainers describe it as a local-first alternative to heavier extraction stacks. For broader format support, the system combines its Rust core with ahead-of-time compiled Apache Tika shared libraries, which allows it to extend parsing coverage while still avoiding traditional server-based overhead. It also supports OCR for images and scanned documents through Tesseract, making it useful for document ingestion pipelines that include image-based or scanned inputs. -
22
Refly
The first open-source agent skills builder
...With a focus on making automation accessible, it provides a visual canvas and low-code components that feel similar to drag-and-drop builders but backed by powerful AI orchestration, memory handling, and integrations with external services. Refly’s approach bridges the gap between workflow ideas and stable, deterministic infrastructure: skills become governed capabilities that can be versioned, shared, and monetized, not just temporary scripts. -
23
IronClaw
IronClaw is OpenClaw inspired but focused on privacy & security
...With dynamic tool building, persistent memory, and background automation, IronClaw is built to securely expand and adapt to your personal and professional workflows. -
24
Tiny CUDA Neural Networks
Lightning fast C++/CUDA neural network framework
...We provide a sample application where an image function (x,y) -> (R,G,B) is learned. The fully fused MLP component of this framework requires a very large amount of shared memory in its default configuration. It will likely only work on an RTX 3090, an RTX 2080 Ti, or high-end enterprise GPUs. Lower-end cards must reduce the n_neurons parameter or use the CutlassMLP (better compatibility but slower) instead. tiny-cuda-nn comes with a PyTorch extension that allows using the fast MLPs and input encodings from within a Python context. ... -
25
LMCache
Supercharge Your LLM with the Fastest KV Cache Layer
LMCache is an extension layer for LLM serving engines that accelerates inference, especially with long contexts, by storing and reusing key-value (KV) attention caches across requests. Instead of rebuilding KV states for repeated or shared text segments, LMCache persists and retrieves them from multiple tiers—GPU memory, CPU DRAM, and local disk—then injects them into subsequent requests to reduce TTFT and increase throughput. Its design supports reuse beyond strict prefix matching and enables sharing across serving instances, improving efficiency under real multi-tenant traffic. ...
