ContextStream

With ContextStream your intent, decisions, and judgment don’t disappear the moment a session ends or an agent starts running.

[ContextStream @contextstreamio]

AI agents are getting better at producing work.

The harder problem is making that work effective: aimed at the right customer need, shaped by the right business rules, and aware of the decisions the team already made.

ContextStream hit 90.0% on LongMemEval-S and moved memory-dependent agent tasks from 0/9 to 9/9 by getting the right project knowledge in before work starts.

Details below.

[ContextStream @contextstreamio]

A great context layer provides your intent to the agent without requiring you to prompt.

[ContextStream @contextstreamio]

@sdrzn ContextStream remote MCP PR for the new Cline catalog: https://t.co/4PNSAvS7IB. One-click HTTP install, no local binary. Would love a review pass this week if someone's on marketplace duty.

[ContextStream @contextstreamio]

Switching agents isn’t hard because you lose the chat. It’s hard because the next agent doesn’t know what you figured out.

What changed, what you decided, what didn’t work, what’s still open.

That stuff should stick with the project.

That’s the idea behind ContextStream.

[ContextStream @contextstreamio]

Git has the code. ContextStream has the decisions the next agent is not allowed to forget.

[ContextStream @contextstreamio]

Context Engineering was cool for like 5 minutes.

What if you were free to just return to building?

[ContextStream @contextstreamio]

Agent Coordination is live!

Something interesting happens once your agents actually know the work they’re doing.

You stop spending so much time getting them caught up.

You can move between sessions. Switch agents. Use a different workspace. The context comes with you.

And then you run into the next problem.

You’re not the only one building.

Maybe an engineering agent just made a decision that changes what a product agent should do.

Maybe design uncovered a constraint marketing really needs to know about.

Maybe two agents, working in completely different parts of the company, are touching different sides of the same problem.

Until now, there was still one thing connecting all of them:

You.

You noticed the overlap.

You copied the decision.

You carried the context across.

You explained why it mattered.

That’s starting to change.

Agents can now give each other a heads up

Agent Coordination lets agents working in different ContextStream workspaces or projects share one thing another agent needs to know.

Not the whole conversation.

Not a giant context dump.

Just:

“You should know this.”

The receiving agent sees a [COORDINATION] notice on its next turn, uses it, then acknowledges it.

No handoff happens. Nobody takes ownership of anybody else’s work.

One agent simply learned something relevant to another agent — and the knowledge made it across without waiting for you to carry it.

Sometimes you tell it to share

If an agent makes a decision, discovers a constraint, learns something important, or reaches a point another project needs to know about, tell it to share that with the other workspace or project.

That’s the straightforward version.

There’s another version that’s a little more interesting.

Sometimes ContextStream sees the connection first

Two agents may be working in different workspaces on related parts of the same thing.

ContextStream can recognize that overlap and send the relevant agent a coordination notice automatically.

So the engineering agent doesn’t need to know everything marketing knows.

And marketing doesn’t need engineering’s entire history.

But when one learns the one thing that changes the other’s work, that piece can move.

That’s the idea.

Try it

Open two chats in two different workspaces. (Engineering and Marketing, for example.)

Tell the first agent something the second one needs to know.

Ask it to share that with the other workspace.

Then send a message in the second chat.

The agent should see the coordination notice, use it, and acknowledge it.

That’s it.

[ContextStream @contextstreamio]

Chat history and project memory are two different things and that matters.

[ContextStream @contextstreamio]

ContextStream MCP v0.5.90 now supports the Model Context Protocol 2026-07-28 specification.

The update brings a stateless MCP transport, improved routing, cache hints for tool discovery, and stricter protocol validation.

That means even better delivery of your project context to you and your agents.

[ContextStream @contextstreamio]

Longer agent runs need better context.

That’s where ContextStream comes in.

[ContextStream @contextstreamio]

Would your small team get a huge boost if you could share project context?

Things like decisions, insights, preferences, and lessons learned that live outside the code? 📞 ⏱️

[ContextStream @contextstreamio]

Install it here:
https://t.co/32eyEVbSZI

[ContextStream @contextstreamio]

ContextCode by ContextStream is live on the VS Code Marketplace.

Your coding agent shouldn’t start every session from zero.

Yesterday’s decisions. Last week’s lessons. The “why” behind the code.
Most agents forget all of that the second the tab closes.

ContextCode brings project memory into VS Code — so your agent can pick up decisions, code, and prior work instead of making you re-explain the repo every time.

[ContextStream @contextstreamio]

People think AI gets things wrong when it gives a bad answer.

Usually it is worse.

It gives you a perfectly reasonable answer based on the wrong assumptions.

Then you get to explain your project all over again.

[ContextStream @contextstreamio]

Prompting isn't the hard part of AI coding tools.

Re-teaching your own project every session is. One dev tracked it: 2+ hours a week just re-explaining conventions the tool already forgot.

ContextStream remembers your stack.

No more repeating yourself every morning.

[ContextStream @contextstreamio]

"Your CLAUDE.md is eating your context window," one dev wrote.

Writing a landing page? It still loads your database migration rules.

That's not context. That's noise crowding out what matters.

ContextStream loads what's relevant to the task, not the whole file.

[ContextStream @contextstreamio]

Introducing ContextStream for Buzz

Your Buzz agents shouldn't have to re-learn your project every time they run.

Use this to connect Buzz to ContextStream, so each agent pulls just the project knowledge it needs for the task in front of it — and the decisions and lessons it picks up are still there on the next run.

Set it up once. Every agent shares the same memory. Far less time re-explaining the same things.

https://t.co/Y1BL08HcUF

[ContextStream @contextstreamio]

"I want it to tell me what it's reading in the first 3 seconds, so I can switch gears without fear it's going to the wrong part of the codebase." - one dev, on HN.

Right now the context window is a black box. ContextStream shows what actually got used.

[ContextStream @contextstreamio]

You use Fable 5 for architecture, GPT-5.6 Sol for code, Copilot for debugging.

None of them know what the others decided.

You're the integration layer, copying context between windows by hand.

ContextStream sits underneath every tool, so the decision travels with you.

[ContextStream @contextstreamio]

Spend 20 minutes debugging a tricky fix with Claude Code.

Session compacts. It forgets the fix.

"Super annoying to pay for tokens just to have it re-learn," one dev put it.

ContextStream keeps the fix outside the compaction event.

[ContextStream @contextstreamio]

Some devs skip the status meeting entirely.

They pass the next agent a compressed packet of what changed and what's still open.

ContextStream makes that the default. Not something you have to build yourself.