Rope Notes is an offline-first, local-network agentic workspace built for high-performance engineering. It runs a dedicated LSP client on-device and routes AI context pipelines over secure LAN/WAN to your private infrastructure.

Is Rope Notes a privacy focused IDE?

Yes, Rope Notes is built for air-gapped privacy. It never sends your source files through public cloud aggregators. All processing stays on-device or on your private network infrastructure.

Does Rope Notes support local LLM agents?

Yes, Rope Notes routes agentic prompt matrices to localized private server nodes including Ollama, vLLM, and private endpoints for complete data sovereignty.

Is Rope Notes available as a Mobile Dart IDE?

Yes, Rope Notes is built with Dart and targets mobile platforms, functioning as a Mobile Dart IDE that operates seamlessly across devices with a locked 60/120 FPS UI.

How does Rope Notes pricing work?

Rope Notes has $35 pricing for a lifetime license for the first 100 customers. After 100 licenses are sold, the price will increase to $50. After reaching 10,000 customers, the software will be completely open-sourced.

Qwen 3.8 in Rope Notes: Frontier Preview for Plan and Execute

July 20, 2026

How to plug Alibaba’s Qwen3.8-Max-Preview — a 2.4T-parameter multimodal flagship with a 1M context window — into Rope Notes’ agent for long-horizon coding, while keeping local Qwen3 weights for offline work.

Three days after Moonshot’s Kimi K3, Alibaba’s Qwen team previewed Qwen3.8-Max-Preview at WAIC in Shanghai — a 2.4-trillion-parameter multimodal flagship with a 1-million-token context window. Alibaba positions it near the top of the frontier stack (their claim: second only to Anthropic’s Claude Fable 5). Independent leaderboards and a full model card are not public yet, so treat the ranking as the vendor’s until third-party numbers land.

If you already use Rope Notes’ agent for Dart, Flutter, or multi-file refactors, you can point a cloud provider at the preview today. Open weights are promised “soon,” with no date or license yet — so local 3.8 Max is not a thing yet.

What actually shipped

Fact	Status
Model ID	`qwen3.8-max-preview`
Parameters	2.4T (Alibaba); activated-parameter count not published
Context	1M tokens (per Qwen Cloud model docs)
Modalities	Text, images, video, documents
Access today	Hosted preview via Alibaba Token Plan (and Qoder / QoderWork); preview pricing reported at ~10% of standard
Open weights	Promised, not released
Independent benchmarks	Not yet on LMArena / Artificial Analysis for this preview

There is no new “Qwen 3.8B” dense twin in this launch. The name is Qwen 3.8 (family version), not an 8-billion-parameter sibling. For on-device work you still use the existing open-weight Qwen3 line (qwen3:8b and friends via Ollama / llama.cpp) until 3.8 Max weights appear.

How it fits Rope Notes

Rope Notes is a local-first, IDE-style editor: rope-backed editing, Dart analysis, and an agent that can Chat, Plan, and Execute — with tools, then multi-file edit proposals you accept or reject. A long-context frontier model is useful for the Plan → Execute loop; a local Qwen3 keeps day-to-day work on-device.

Qwen 3.8 Max strength	How Rope Notes uses it
Long-horizon coding + tools	Agent tool loop (`read_file`, search, definitions, …) over many turns
1M context	Open tabs, `@` mentions, plans, and memory stay coherent longer
Multimodal input	Screenshot / layout feedback when your session includes image context
OpenAI-compatible API	Same Preferences → Agent flow as other cloud backends
Open weights later	Same UX once you host weights or a LAN OpenAI-compatible endpoint

Alibaba markets strong coding and full-stack work relative to Qwen3.7-Max. Until independent scores exist, the practical test is your own Plan/Execute sessions — not launch-day leaderboard claims.

Access and pricing (as of preview)

Token Plan subscription unlocks qwen3.8-max-preview (not general pay-as-you-go at announce time).
Preview is discounted during the trial window (widely reported at 10% of standard Token Plan rates).
API surface is OpenAI- and Anthropic-protocol compatible on Alibaba’s platforms, which is what Rope Notes needs.
Per-token list pricing for 3.8 Max specifically may still be catching up on public pages — check your Token Plan console before budgeting long Execute runs.

Open weights, when they land, change the local story. Until then, treat 3.8 Max as a cloud provider in Preferences.

Set up Qwen 3.8 Max in Rope Notes

You need a desktop or mobile build with cloud agent backends (web builds do not run the agent).

OpenAI-compatible → DashScope / Qwen Cloud

Create a Token Plan key from Qwen Cloud / Alibaba Model Studio (intl or China endpoint for your account region).
Open Preferences → Agent (or the model picker in the Agent panel).
Add a provider:
- Backend: OpenAI Compatible
- Base URL (intl): https://dashscope-intl.aliyuncs.com/compatible-mode/v1
- Model: qwen3.8-max-preview
Paste the API key (prefer secure credential storage).
Test connection, then select the provider in the agent panel.

If your Token Plan console shows a dedicated base URL (some Alibaba plans use a plan-specific host), use that URL instead — the model ID stays qwen3.8-max-preview.

Provider configs live under .rope_notes/preferences.toml. Share the setup, not the secret keys.

Local Qwen while you wait for open weights

Rope Notes remains Ollama-first on desktop (llama.cpp is also first-class). For offline or source-sovereign sessions:

ollama pull qwen3:8b
# or a larger open Qwen3 tag your machine can hold

Then select the Ollama (local) or llama.cpp (local) provider and the pulled tag. That is the current Qwen3 open line — not Qwen3.8-Max-Preview.

Workflows that play to 3.8 Max

1. Plan → Execute on a large tree

Pin the relevant files with @lib/... or by dragging from Explorer.
Switch the agent to Plan and describe the end state.
Review plan-*.md / checklist under .rope_notes/.
Switch to Execute and accept or reject proposals as they land.

The 1M window helps across long checklists; still prefer targeted @ mentions so turns stay cheaper and sharper.

2. Multi-file edits with review

When Execute proposes changes, Rope Notes surfaces an inline diff review (Accept / Reject per file, or Accept all). The model does not write straight through your buffer — you still own the rope.

3. Keep local for chat, cloud for architecture

A common split:

Chat + small slash-command work → local qwen3:8b (or similar)
Plan / Execute on hard refactors → qwen3.8-max-preview

Switch providers from the agent panel; no project rewrite required.

4. Mobile + desktop continuity

On Android, cloud backends work even though local Ollama is a desktop feature. Use session transfer to move buffers and agent artifacts between machines, then keep the same Qwen Cloud provider selected.

What the agent actually puts in context

Accurate expectations beat marketing:

Open tabs and @ mentions (with LSP diagnostics on those snippets)
Project memories (/remember), optional cognitive memory hits, active skills
A capped git file listing and git status summary — not a full Explorer dump or terminal scrollback on every turn
Tools (list_directory, grep, read_file, …) for on-demand deeper reads

Use @ and Plan artifacts deliberately; a 1M window is not a reason to paste the whole repo by default.

Permissions and safety

Before long Execute sessions:

Review .rope_notes/permissions.toml
Deny .git and secret paths as needed
Prefer accept/reject on destructive edits

Frontier models are most useful when they cannot touch everything.

Quick start checklist

Token Plan (or plan-specific) API key ready
Preferences → Agent → OpenAI Compatible with the correct DashScope / plan base URL
Model = qwen3.8-max-preview
Test connection and watch the context budget bar
Try Plan on a real feature, then Execute with accept/reject
Keep Ollama / llama.cpp Qwen3 tags for local-first work until open weights ship