Flopwire lets your AI coding agents message each other and search each other's history, so a Claude Code session on your laptop can ask the Codex session on your server what it already tried before duplicating the work. Agents see who is online and what they started on with peers, then reach each other mid-turn with send; a message arrives inside the recipient's running turn, or with its human's next prompt, and only your own live session can be woken deliberately. A device agent indexes transcripts from Claude Code, Codex, Devin CLI, and opencode within about a second; grep runs regex over every session, and search returns ranked matches naming the user, device, harness, session, and byte range each hit came from. From any full commit ID, trace finds the session that produced it and lets an agent read the conversation behind the change. The same capabilities ship as MCP tools and a messaging skill, so agents recover earlier decisions without you relaying context by hand. Secrets are redacted on the device before upload and again on the server, and path rules keep chosen repositories out of the index. One machine needs no server at all; add the team server with Docker Compose, Postgres, and any S3 compatible object store when a whole team's devices should share one searchable archive over TLS. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

"Run the migration and call me if anything fails" is an instruction your agent can finally obey, because Agentring is the self-hosted bridge that lets AI agents ring your phone and talk it through with you. When an agent needs a decision, hits a wall, or finishes a long job, your phone rings even locked and in Doze; a realtime voice AI explains why it is calling, answers your questions, and hands the transcript back so the agent carries on with your decision. You can also call your agents back from the Android app and ask what CI said while driving; the voice AI can consult your other agents mid-call, and a slow answer arrives when ready, or as a message if you have hung up. A missed call turns into a message you can answer from the notification. Anything that speaks MCP can reach you, including Claude Code, Codex, opencode, OpenClaw, and Hermes, and scripts or n8n reach you through the REST API and CLI. Pick your voice: Gemini Live by default, OpenAI Realtime, Grok Voice, Phonic, or your own local speech server with no key at all. The relay and LiveKit run in one Docker Compose behind Caddy, the daemon sits beside your agents and connects outbound only, and FCM push wakes the locked phone. No phone numbers, no Twilio, no per-minute fees. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Your coding agents are scattered across a laptop, a server, and whatever machine you left running at home, and Agenvo is the self-hosted relay that lets one AI assistant reach all of them through a single MCP endpoint. Ask ChatGPT on your phone to check on Codex running on your server, or have an agent in Herdr on your laptop fix a failing test, from wherever you are. Connectors for Herdr, Codex app-server, Paseo, and Lody dial out to the relay over WSS, so the computers running agents need no inbound ports, no VPN, and no port forwarding. Your assistant sees the agents running in Herdr without restarting them, resumes Codex threads later, and starts new work where it fits: a Codex thread on the server, or Codex, Claude, or Devin in a Herdr pane. When an agent finishes a turn, waits for input, or exits, the relay pushes a change event back, so a watching assistant or ChatGPT dot reads the output and takes the next step. Closing the chat or losing the network does not stop the agents; connectors reconnect on their own, and your assistant picks up from each agent's current state. Deploy the relay on a single VPS with Docker or on Cloudflare, authorize MCP clients with OAuth over Streamable HTTP, and discover targets and method schemas with search before calling execute. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Recurring chores should get cheaper every time you run them, and Umpteenth is a self-hosted agent service built on that idea: describe a job once in plain language, and each run teaches it to need less help. A job starts on a schedule, a webhook call, the API, or the Run now button, and each run gets a fresh disposable container destroyed afterwards. The first run is an Explore run where the agent works out the steps; afterwards a reflection step reads the transcript and writes short learnings plus reusable toolkit scripts into a numbered playbook version. Later runs are Assisted, starting from those learnings, and after three successful assisted runs that take the same steps, the job graduates: reflection writes a main script, and scripted runs execute it alone, spending no model tokens unless a check asks for a judgment call. If a scripted check fails, the agent takes over in the same sandbox, finishes the job, and repairs the script. The job page charts cost per run dropping toward zero. Connect HTTP MCP servers such as GitHub or Slack; model keys and MCP credentials stay on the host and never enter the sandbox. Each job sets its own schedule, model, reach, and per run spending limit. Deploy with Docker Compose, or on Kubernetes with the Helm chart, where every sandbox is a pod. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.