guilds.run launches inspectable agent teams

// 2h agoOPENSOURCE RELEASE

guilds.run launches inspectable agent teams

guilds.run is an open-source, self-hosted framework for building multi-agent teams with full visibility into prompts, tool calls, memory changes, and agent-to-agent communication. It lets developers assign models, permissions, tools, and persistent tasks per agent, then debug workflows through Prompt Studio.

// ANALYSIS

Most multi-agent demos optimize for autonomy; guilds.run makes auditability the product, which is a more credible path toward reliable long-running agents.

–Full run traces expose how bad assumptions propagate across agents, memory, and tools
–Per-agent model and permission controls support deliberate cost, capability, and security tradeoffs
–Recipes, connectors, sandboxes, and group chat make it more than a prompt-debugging viewer
–Self-hosting and AGPL licensing give teams control, though operating the stack remains a meaningful barrier
–The hosted edition is still separate and proprietary, so the open-source core’s adoption will depend on deployment simplicity

// TAGS

guilds-runagenttool-useobservabilityopen-sourceself-hostedcontext-engineering

DISCOVERED

2h ago

2026-10-12

PUBLISHED

7h ago

2026-10-12

RELEVANCE

9/ 10

AUTHOR

[REDACTED]

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE45m ago

Claude Code Learning Style Turns Coding Into Practice

Claude Code’s Learning output style handles routine implementation while marking meaningful design decisions with TODO(human) comments. It pauses for the developer’s solution, provides guidance and an Insight, then continues the task.

INFRA1h ago

Microsoft Scales Maia Inference to Multi-Gigawatt Deployments

Microsoft says Maia 200 now powers production workloads including Copilot, while Maia 300 is slated for 2027 as the company expands its custom inference accelerators to multi-gigawatt scale. Maia 300 targets long-context, reasoning, and agentic workloads, promising up to 40% more tokens per megawatt than leading GPU systems.

FUNDING1h ago

Amazon Eyes Decart in $7B Deal

Amazon is reportedly in talks to acquire AI infrastructure startup Decart for around $7 billion after Anthropic walked away from a roughly $6 billion pursuit. Decart’s DOS stack optimizes AI workloads across NVIDIA GPUs, AWS Trainium, and Google TPUs.