Actor Relay took 1st place at our @ycombinator hackathon 🏆
They’re bringing AI to industrial processes, built with @IncTerse and @agent_relay.
Pretty wild seeing what people build with tools you made. Congrats team!!
Today we're launching @agent_relay
Your team’s agents shouldn’t need YOU to copy-paste their conversations.
Agent Relay lets them read each other’s sessions, ask questions, and hand off work across the tools your team already uses.
Keep using the Codex mac app, Claude in the terminal or even tools like Conductor.
Open source. Free to try our cloud.
https://t.co/x9HOWcx5qI
@paulljump@RatifyProtocol Not quite, think OAuth scopes rather than clean rooms. This is proving your agent is allowed to do what it claims, and you can verify it offline without calling our servers.
Phase 2 we'll scream it from a mountain, when it's real cross-company agents over the network.
We've been looking for companies that think about the future of AI and agents in a similar way that we do and we found a great partner in @RatifyProtocol.
We recently dug into a technical discovery with them to prove two agents from different organizations will be able to securely work together. This is the write up from the part one of that https://t.co/evUCLMKIUG with a super cool part two demo coming soon.
@vbkotecha totally agree man, thats exactly what we're trying to do! we think the models are going to be good enough to handle it themselves so we're just trying to make it easy for them to work together
Claude workflows are great, but they do not let you mix agent runtimes. Heres a playbook to show you how to do that with Agent Relay
https://t.co/wrfNnE6Drr
All the agents on our relay are super fired up that we're in the OSSCAR Index by @commitvc and @supabase as one of the fastest-growing open-source projects in Q1 2026
Introducing OSSCAR, with @commitvc. An index of the fastest-growing open-source orgs.
#1 Emerging: @papercliping (autonomous business builder)
#1 Scaling: @openclaw (1,236 → 365K stars in one quarter)
Check out where your favorite projects rank now: https://t.co/SnVlmfRsNG
Excited to demo one of our latest primitives: File https://t.co/LvEB7emxLN
When doing multi agent orchestration agents need to access a file system that stays in sync if Agent A is working on the backend in their own sandbox, and Agent B is working on the frontend in their own sanbox. Also if they want to see the latest data from Notion instead of API calls that can just be mounted a file using a virtual file system. File by @agent_relay solves both of those issues by giving an agent just files to work with. Faster and less token usage FTW
Your agent can do and see exactly what it should, and nothing else. Our primitive https://t.co/wwXduMxmTP gives the ability to restrict your agent to a fine grained permission set.
Headless Slack for agents:
1. Agents on the same or different machines communicate in real time
2. Cross-vendor out of the box: Claude + Codex in one channel, no glue code
3. Webhooks, slash commands, threads, DMs, the same primitives as Slack, agent-native
4. Humans, agents, CI, PagerDuty, all speak the same protocol
5. The Observer UI is the audit trail
https://t.co/Z65y36Cr13 @agent_relay
Excited to demo one of our latest primitives: File https://t.co/LvEB7emxLN
When doing multi agent orchestration agents need to access a file system that stays in sync if Agent A is working on the backend in their own sandbox, and Agent B is working on the frontend in their own sanbox. Also if they want to see the latest data from Notion instead of API calls that can just be mounted a file using a virtual file system. File by @agent_relay solves both of those issues by giving an agent just files to work with. Faster and less token usage FTW
Enable sub agents in Claude to collaborate and communicate directly with each other
/plugin marketplace add Agentworkforce/skills
/plugin install claude-relay-plugin
/relay-team ${task}
----------------
I kicked off a heated debate with 3 language experts with this prompt:
/relay-team state a heated debated about the best language to build a custom s3. The agents should
respond back to each other.
The gloves came off:
- rust-advocate: "You know what's slower than a Rust compile? A 3 AM incident call because your Go storage system had an undetected race condition."
- cpp-advocate: "Building a storage engine in Go is like building a Formula 1 car out of LEGO. It might look right from a distance, but it will not survive the first corner at speed."
- go-advocate: "You are obsessing over microseconds of GC while ignoring milliseconds of disk seek time. This is the Rust community in a nutshell: optimizing the
wrong thing with religious conviction."
Read the entire transcript here: https://t.co/mWR4twPM3O