Three months after our very first public alpha — and just one week after shipping Alpha v0.0.1 — we’re back with Alpha v0.0.2 (published Aug 17, 2025). This tag is all about faster prototyping, stronger security, and smoother execution. (GitHub)
⚠️ Alpha quality: still not production — please test, break, and report. (GitHub)
What’s new
Rapid Prototyping Mode (Layered Nodes)
The biggest addition in this release is Rapid Prototyping mode. You can now drop in placeholder nodes to sketch workflows before wiring them to real technical implementations.
Instead of switching between “business idea” and “technical solution,” you can now combine both in one flow. Early brainstorming is smoother, and sharing concepts with teammates is easier.
Here’s a short demo:
This lays the foundation for fast idea-to-execution pipelines: start with abstract nodes, then refine them into fully operational ones as the project matures.
New Models: Qwen 3 & Qwen 2.5 VL
We’ve added support for two major new models:
- Qwen 3 → a strong reasoning-first language model, tuned for logic, analysis, and code-heavy workflows.
- Qwen 2.5 VL → a multimodal vision-language model that handles images + text together, making it perfect for visual automation, annotation, and AI agents that need to “see” and “read.”
These models join our growing library, giving you broader options for experimentation and production-like testing.
Lockable Layers
A small but highly requested feature: you can now lock layers. No more accidental nudges when presenting or collaborating — lock them in place once your structure is set, and focus on the logic.
New E-Mail Nodes (SMTP + IMAP)
We’re bringing automation straight to your inbox. With the new generic SMTP and IMAP nodes, you can now:
- Read your mails directly in a flow
- Organize mails into folders automatically
- Manage your folder structure
- Create drafts for review or scheduled sending
- Send messages directly from Flow-Like
- …and much more!
Everything is ready for mailbox automation — whether you’re building personal workflows or enterprise-scale mail processing pipelines.
Secure Offline Sharing (xChaCha20 + BLAKE3-HMAC)
For teams that need to share flows offline or in air-gapped environments, we’ve built a secure sharing mechanism powered by modern cryptography:
- xChaCha20 → a stream cipher designed for performance and safety, with a 192-bit nonce that eliminates the risks of nonce reuse.
- BLAKE3 (HMAC) → a fast, parallelizable hashing algorithm used here for authentication and integrity checks.
The combination means you get enterprise-grade encryption without sacrificing speed, ensuring your flows remain private and tamper-proof.
Execution Fixes: Nested + Parallel
We’ve smoothed out execution logic across the board:
- Nested Parallel Execution now behaves as expected.
- Parallel Model Execution allows for parallel model and embedding model invocation.
Complex orchestration is now much more reliable.
Thank you, community & contributors
This release was heavily shaped by your feedback from v0.0.1. Every report, idea, and suggestion helps make Flow-Like stronger. Please keep them coming. 💙 (GitHub)
Get it & try it
Grab installers and artifacts from the Alpha v0.0.2 release page — then open an issue with repro steps if something breaks. We’re still reading everything. (GitHub)
What’s next
- Richer typing across nodes and data edges
- Built-in evaluation tools for flows
- Governance & observability for enterprise workflows
If you’re experimenting with us this early, you’re helping shape the foundation of Flow-Like. Thanks for being here!