Mercury Agent teases biggest update ever

// 2h agoPRODUCT UPDATE

Mercury Agent teases biggest update ever

Cosmic Stack Labs' Mercury Agent, an open-source AI agent designed to function as an "always-on" personal assistant with structured persistent memory, permission-hardened tools, and multi-channel access (CLI/Telegram), has teased its largest update to date. The update was announced via their official X account, hinting at significant new features and capabilities for the autonomous assistant.

// ANALYSIS

While "soul-driven" and "always-on" sound highly ambitious, the true test for Mercury Agent's upcoming update will be whether its permission-hardened tool access and daemon-mode stability can actually outperform more established agent frameworks.

* Cosmic Stack Labs is aiming to carve out a niche for Mercury Agent by focusing heavily on persistent user memory and multi-channel accessibility (Telegram/CLI).

* The announcement is a teaser ("Soon"), so the concrete details of the update remain under wraps, raising expectations for their next release.

* Daemon mode and budget control are highly practical features that address real-world deployment challenges of long-running autonomous agents.

// TAGS

agentopen-sourcecosmic-stack-labschatbottelegramdeveloper-tools

DISCOVERED

2h ago

2026-07-18

PUBLISHED

2h ago

2026-07-18

RELEVANCE

6/ 10

AUTHOR

mercury__agent

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

MODEL19m ago

OpenRouter adds nine new AI models

Unified API provider OpenRouter has added nine major new AI models to its platform, highlighted by Moonshot AI's Kimi K3, Meta AI's Muse Spark 1.1, and Thinking Machines Lab's Inkling. The additions provide developers with immediate API access to these frontier systems for tasks ranging from long-horizon coding and tool use to multimodal reasoning.

UPDATE1h ago

Tesana automates character weapon rigging

Tesana AI has rolled out an engine upgrade that automates character weapon and item attachments, bypassing the tedious manual rigging process. By automatically handling grip points and alignment, the engine allows developers to speed up asset importing and focus on core game design.

BENCHMARK1h ago

GLM-5.2 matches closed models on cyber tasks

The UK AI Security Institute (AISI) has released evaluation results from testing leading open-weight AI models against closed frontier systems on practical cyber work, such as vulnerability research, reverse engineering, exploitation, and multi-step network attacks. The benchmark results indicate that the performance gap between open-weight and closed-weight models is shrinking rapidly, with Z.ai's open-weight GLM-5.2 matching the cyber capabilities of closed frontier models released just four to seven months prior.