Mercury Agent adds conscious, subconscious memory

// 1h agoOPENSOURCE RELEASE

Mercury Agent adds conscious, subconscious memory

Mercury Agent by Cosmic Stack introduces a dual-layer memory system and permission-gated shell access to transition AI agents from "stateless interns" to persistent system daemons. The update focuses on long-term context retention and safe autonomous execution for terminal-based workflows.

// ANALYSIS

Mercury’s "soul-driven" architecture is a direct response to the context window limitations and reliability gaps in current autonomous agents.

–The conscious/subconscious split allows for high-speed reasoning on active tasks while maintaining a deep, searchable history of all past interactions via local SQLite storage.
–Permission-gated actions provide a critical safety layer, requiring explicit user approval for high-impact commands like file deletions or sensitive API calls.
–Token-aware execution and background daemon capability enable 24/7 autonomous workflows that can be managed via CLI or Telegram.
–By prioritizing a local-first, open-source stack, Mercury offers a high-privacy alternative to cloud-locked competitors like Devin or Manus.

// TAGS

mercury-agentagentagent-memoryai-codingclitool-useopen-sourcerag

DISCOVERED

1h ago

2026-05-24

PUBLISHED

2h ago

2026-05-24

RELEVANCE

9/ 10

AUTHOR

mercury__agent

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

VIDEO37m ago

Resend turns Claude visibility into strategy

In this Better Stack episode, Resend frames AI discovery as a deliberate growth channel: they want to show up when developers ask Claude for an email API, and they back that up with structured content, clear Q&A sections, and early adoption of LLM-friendly tooling. The broader takeaway is that modern developer products can shape how they are discovered by both search engines and AI assistants if they make their product legible to machines as well as people.

NEWS1h ago

Anthropic readies Mythos 1 for Claude Code

Anthropic's "Mythos-class" frontier model tier surfaces in internal app traces, signaling a massive leap in autonomous coding and defensive cybersecurity capabilities. The model is currently restricted to select enterprise partners through Project Glasswing to ensure robust safeguards before a wider release.

NEWS1h ago

GPT-5.6 leaks point to June launch

Leaks point to a June launch for GPT-5.6, featuring "active inference" reasoning and a 1.5M context window. The model reportedly fixes reward-model contamination issues identified in the April GPT-5.5 release.