bytedance/deer-flow

Prefer upstream unless you specifically want a minimally modified snapshot. This fork adds no visible capabilities, while still lagging the latest upstream work.

Choose this fork only if you need a near-vanilla DeerFlow 2.0 snapshot. If you want active features, fixes, or security updates, upstream is the better choice.

Choose this fork if you want a minimal UI privacy/simplification tweak and can tolerate being far behind upstream. If you need the newest DeerFlow integrations and fixes, upstream is the better fit.

Choose this fork if you want DeerFlow plus deeper embedding, memory-management, and integration changes, and you are comfortable maintaining a materially diverged codebase. Choose upstream if you value current fixes, lower maintenance risk, and broader feature parity.

Choose this fork if you want a more RAG/knowledge-base-centered DeerFlow with custom infrastructure support and you are comfortable maintaining a divergent codebase. Choose upstream if you want the latest active work, security hardening, and broader platform integrations.

Choose this fork if you want a more customized, replay-oriented DeerFlow variant and are comfortable owning merge debt. Choose upstream if you want the newest agent, memory, gateway, and sandbox work with lower maintenance risk.

Choose this fork only if you specifically want a near-upstream copy. For most adopters, upstream is the better choice because this fork adds no visible features and is already behind by a few commits.

Prefer upstream unless you specifically want this fork as a starting point or workspace. It offers no visible added capability, and it is slightly behind the latest upstream feature work.

Choose this fork only if the `/mnt/skills` local-sandbox access change solves a concrete blocker for you. Otherwise, upstream is the better default because this fork is far behind and adds only a small local customization.