Shane Thomas retweeted
We completed our incident report + timeline for Wednesday's platform outage + the security breach at our underlying infra provider Railway:
mastra.ai/blog/platform-inci…
cool to see that Halo uses mastra!
New article on how we built Halo 😇
A couple of decisions:
- Halo uses the same MCP server, skill and docs as external agents accessing Langfuse, so every new Halo functionality is immediately available to everyone.
- The agent loop runs separate from its sandbox.
- We use Langfuse to observe and improve Halo.
Read the full article here: langfuse.com/blog/2026-10-21…
Shane Thomas retweeted
Most LLMs get "context rot": recall decreases as context window approaches the context limit.
We wondered: do the new "System 1" decision models get context rot? So we tested Jev, Clef, and Luna Decisions across 400 questions.
Turns out decision models get context rot, too.
Shane Thomas retweeted
Mastra's platform is hosted on Railway and thus Mastra platform users are affected by the ongoing Railway security incident.
Comms here as things evolve.
Shane Thomas retweeted
Today we're launching support for the latest MCP spec (MCP 2.0) in Mastra!
You can now scale servers without managing sessions and route requests using HTTP headers:
Shane Thomas retweeted
We’ve worked with hundreds of teams shipping Mastra agents into production. Now, it's your turn.
We're introducing Mastra Launch and offering hands-on FDE support to get your agent live within 90 days.
Let's chat: calendly.com/mastra-team/lau…
Last week on Agents Hour:
- espionage at OpenAI,
- Mistral's chonky new model
- Haiku 5.5 and Gemini 4 Argon
go subscribe to Agents Hour on Youtube, Spotify, or your fav podcast app: youtube.com/@mastra-ai
Shane Thomas retweeted
Presenting Livebase, a demo app showing how to build collaborative human <> agent systems on a @neondatabase + @mastra stack.
Shane Thomas retweeted
Today we're incredibly excited to launch trace queries!
The point of collecting observability data is to analyze it and improve your agent.
So now you (or your coding agent) can query by fields like time, primitive name, agent/workflow ID, status, score, and feedback:
Shane Thomas retweeted
21 months later
same restaurant
okay the gatsby gang was back together this week in SF for a @mastra onsite
so this week's changelog is pretty epic 👇
Shane Thomas retweeted
Today we turned off external PRs for Mastra
We noticed we’re almost always closing PRs and having our factory do them anyway
Instead we’re encouraging contributors to open issues with detailed repros
Shane Thomas retweeted
Today we're excited to launch trace imports!
Export traces from Langfuse to the Mastra platform:
OpenAI Caught a Spy! Plus Mistral Le Chonk & Claude Haiku 5.5 xitter.cf/i/broadcasts/1RJjpbZoR…
Agents write code faster than CI can check it.
GitHub Actions minutes are up 4x. Anthropic's CI job volume grew 25x in 6 months.
@dbworku thinks CI as we know it doesn't survive.
subscribe to Agents Hour on YouTube or your favorite podcast app open.spotify.com/show/0hR0yV…
Shane Thomas retweeted
Today, we're launching Sazabi version 1.0.
The reality of shipping software as a team in the age of AI is actually crazy.
Agents merging their own PRs. Hundreds of deployments per day. A codebase that is constantly changing. Junior developers with 10x or 100x leverage. 😵💫
I spent the last year building a safety harness for teams that ship with AI. It's called @sazabi.
The premise is simple:
1. You send us logs.
2. We detect anomalies, investigate them, and send you back rich, actionable alerts about real problems in your app.
Sazabi helps you find issues before your users notice.
No monitors or noisy alerts. Just pure signal.
And today we're making it available for anyone to use.