The team says it right away: we can’t consider this new version as ready to replace the main client in C, far from it. But the work is progressing well, and Meaning 0.1.0 now has high-level APIs stable enough for others to try.
Interested people can embark them on experimental projects to measure the benefits and identify blockages. However, they are warned: Tor cannot guarantee that no breakout will occur in the coming months. Before adding that it will not be without a good reason.
Among the biggest changes, the API Builder now supports the creation of unbootstrapped and bootstrap-on-demand Tor instances. The Error API has been replaced. The new one is much less verbose than the old one, which tended to reveal many lower-level APIs.
the path to version 1.0.0 will be intense. The team plans to deliver it in September and must, in just six months, bring it to full parity with the C client in terms of network performance, CPU usage, resilience and security features.