EN
Back to the archive

The archive · AI & Models · Product decision · 2025–2026

RightNow's OpenFang bet: a Rust 'agent OS' for scheduled work — 17.9K stars

RightNow's MIT open-source OpenFang is a Rust 'agent operating system' whose seven autonomous Hands run on schedules; it hit 15.9K GitHub stars in Q1 2026.

RightNow

The betThat agents are an operating system, not a chat: scheduled Hands run 24/7 - researching and generating leads - sandboxed, audited and approval-gated instead of prompted.Live

What the business is

OpenFang is RightNow's open-source 'Agent Operating System' built from scratch in Rust: roughly 137K lines across 14 crates with 1,767+ tests, compiling to a single about-32MB binary. Instead of waiting for chat input, its seven bundled Hands (Clip, Lead, Collector, Predictor, Researcher, X account manager, Browser) run autonomously on schedules, building knowledge graphs, monitoring targets, generating leads and reporting to a local dashboard. The system includes 40 channel adapters, 27 LLM providers, 53 built-in tools plus MCP and A2A support, 16 security systems (including a WASM dual-metered sandbox and Merkle hash-chain audit trail), a WhatsApp Web gateway, an OpenAI-compatible API, a migration engine from OpenClaw, and a Tauri desktop app. It is MIT-licensed and maintained by Jaber, founder of RightNow, which Runa Capital's ROSS Index describes as founded 2025 in Amman, Jordan.

How it started

Jaber, founder of RightNow, built OpenFang from scratch in Rust and released v0.1.0 in February 2026 as a public first version. The positioning was deliberately contrarian: not a chatbot framework, not a Python wrapper around an LLM, not a multi-agent orchestrator, but a full 'agent operating system' whose agents run on schedules 24/7 - waking up, doing multi-phase work, and delivering results to Telegram or a dashboard without the user typing each prompt.

What happened

Traction compounded through spring 2026: the repository passed 15.9K stars by the end of Q1 and made Runa Capital's ROSS top-20, then reached 17.9K stars by the Q2 observation end of 2026-05-27. ROSS also reported RightNow had raised $0.54M from Y Combinator and Maza Ventures. The repo moved fast - v0.5.10 by April 2026 - while DiarioBitcoin and other outlets covered the launch, and feature breadth grew to seven bundled Hands, 40 channel adapters, 27 LLM providers, 53 tools and 16 security systems.

How it ended up

As of September 2026 OpenFang remains open source and pre-1.0: the repository says it is 'feature complete but still pre-1.0', recommends pinning to a specific commit for production use until v1.0, and continues to document the full agent-OS architecture. The star growth validated developer appetite for scheduled autonomous agents, while monetization, pre-1.0 stability and the sheer breadth of the platform remain open questions.

Background

RightNow's OpenFang, first released as v0.1.0 in February 2026, is an open-source 'Agent Operating System' built from scratch in Rust: about 137K lines across 14 crates, 1,767+ tests, and a single roughly-32MB binary - explicitly not a chatbot framework or LLM wrapper.

Its seven bundled Hands (Clip, Lead, Collector, Predictor, Researcher, X account manager, Browser) run autonomously on schedules - monitoring targets, building knowledge graphs, generating qualified leads, cutting and posting videos, and reporting to a dashboard - with 40 channel adapters, 27 LLM providers, 53 tools, and 16 security systems including a WASM sandbox, Merkle audit trail and approval gates.

Traction: Runa Capital's ROSS Index ranked the repo in its top-20 for Q1 2026 at 15.9K GitHub stars (as of 2026-03-31) and Q2 2026 at 17.9K stars (observation end 2026-05-27), and reported $0.54M raised from Y Combinator and Maza Ventures; DiarioBitcoin covered the February 2026 launch.

As of September 2026 the project is MIT-licensed, maintained by RightNow founder Jaber, and still pre-1.0 - described as 'feature complete' with breaking changes expected until v1.0 - leaving monetization and stability as the open questions behind the viral growth.

What has to be true

  • OpenFang made a paradigm concrete - scheduled, self-running Hands that work 24/7 versus chat-driven assistants that wait for prompts - and that distinction spread quickly on GitHub.
  • It led with engineering credibility a skeptical developer audience could verify: Rust, one binary, 1,767+ tests, 16 security layers, 40 channel adapters and a migration path from OpenClaw.
  • Timing aligned: agent tools dominated GitHub's fastest-growing lists in early 2026, and ROSS reported YC and Maza Ventures backing as the repo grew from 15.9K to 17.9K stars across Q1-Q2.
  • The bet validated demand for autonomous scheduled agents, but the platform's scope - still pre-1.0 with no disclosed revenue - leaves the business case unproven.

What can be applied

Differentiate on the job, not the model: scheduled, sandboxed Hands with channel adapters beat chat sessions - but the breadth that earned 15.9K stars in Q1 2026 leaves OpenFang pre-1.0.

Aftermath

As of September 4, 2026, OpenFang remained MIT-licensed and pre-1.0 under RightNow-AI, maintained by founder Jaber, with production users told to pin a commit until v1.0. Its Q1-Q2 2026 run - 15.9K GitHub stars (2026-03-31), 17.9K by the ROSS Q2 observation end of 2026-05-27, $0.54M raised from Y Combinator and Maza Ventures - made it one of the fastest-growing agent projects of the period; DiarioBitcoin covered the February launch. Seven Hands, 40 channel adapters and 16 security systems remain documented, while monetization, v1.0 stability and scope stay open.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases