
Magic Context is a context and memory plugin for developers working with coding agents in OpenCode, Pi, and OMP. Part of CortexKit, it keeps project knowledge across sessions and handles history compression in the background so context management doesn't interrupt ongoing work.
As it summarizes conversations, it saves decisions, constraints, conventions, and configuration values as project memories. Those memories carry into future sessions and across the supported agents, reducing the need to explain the same project rules again. The agent can also record or delete memories directly.
Older conversations become chronological summaries with importance scores. Their level of detail adjusts to the model's context window, keeping the active context smaller while retaining the thread of earlier work. The summarization model is separate from the main coding model: you can use a fully local model for that background task while keeping a different model for coding. That choice applies to summarization, rather than establishing that the whole workflow runs offline.
An optional background agent checks stored memories against the codebase, corrects stale facts, merges duplicates, and records lessons from corrections. Relevant memories surface automatically each turn. When an agent needs more detail, it can search project memories, past conversations, git commits, notes, and primers through semantic search with a full-text fallback.
Magic Context is open source under the MIT license.
Claim this page with an email at discord.gg. Magic Context gets the verified badge, and you can upgrade the listing to be featured on localhosted. Proud to be listed? Put our badge on your site.
Want more people to find Magic Context?Promote it
Something wrong or outdated on this page?
3.7KUpdated 5 months agoApache-2.0
#Agent Skills#Code execution#Context compression
Acontext turns AI agent sessions into Markdown skill files that agents can reuse on later tasks. It's for developers who want agents to retain successful approaches, past mistakes, and user preferences in memory they can inspect and correct. The full stack can run on your own infrastructure, and a hosted cloud service is also available. It's open source under Apache 2.0.
2.6KUpdated 2 days agoApache-2.0
#Distributed execution#Guardrails#Human approval
1.6KUpdated 6 days agoApache-2.0
macOS · Windows · Linux#Agent Skills#Hugging Face integration#Human approval
29.1KUpdated 10 hours agoApache-2.0
macOS · Windows · Linux · Docker · Web#Agent Skills#Git integration#Hybrid search
1.1KUpdated 20 hours agoAGPL-3.0
macOS · Web#Agent Client Protocol#Agent Skills#Git integration
8.7KUpdated 20 hours agoMIT
macOS · Windows · Linux · Docker · Web#Git integration#Hybrid search#Knowledge graphs
AgentField is a self-hosted, open source platform for developers building AI agents that need to work together under shared access rules. It runs in your own infrastructure and combines agent execution with identity, policy enforcement and audit records. The project uses the Apache 2.0 license.
Agentlas OS is a local-first AI agent framework that assembles specialist teams around a task. It runs inside Claude Code, Codex, Cursor and Gemini and is open source under Apache 2.0. Its roles cover software development and work such as insurance analysis, M&A diligence and travel planning.
agentmemory is a self-hosted memory server for developers who want their coding agents to retain project decisions, past fixes and preferences across sessions. It captures activity through agent hooks and retrieves relevant context for later work. Multiple agents can share the same memory, so switching between Claude Code, Cursor and Codex CLI doesn't mean starting with an empty history.
AgentRQ is a self-hosted task manager for people supervising AI agents on their own computers. Agents can ask for decisions, report progress, and receive tasks while you follow their work and reply from a browser, phone, or desktop app. It also offers a hosted service; agents run on your machines, while the task dashboard runs on your own server or the hosted service.
ai-memory keeps project knowledge on a server you control so developers can switch coding assistants or machines without losing context. It connects Claude Code, Codex, Cursor, Gemini CLI and OpenCode to shared memory, including unfinished work, failed approaches and open questions.