qwen-code adds scoped external memory import, starting with Codex migration
callmeYe · ghdev · 2026-10-15
PR #14281 in QwenLM/qwen-code (+6130/-35, 46 files) adds a Qwen-owned external memory import service for CLI and cloud integrations, starting with Codex.
- Flow: clients submit relative Markdown paths and contents, inspect a deterministic preview, explicitly select sections and destination scope, then start durable background consolidation using the workspace's configured model; results enter native memory search/fetch/recall/forget
- Design: cloud runtimes can't read the user's laptop filesystem, and per-bullet conversion loses historical context without reconciling existing memories — so the server owns destination paths, validates workspace trust, never falls back to the primary runtime, and retains only selected source snapshots
- Reliability: journaled publication, idempotency, cancellation and retry protect existing/edited/forgotten memories; bounded per-project batches with a same-scope metadata index and durable call/token reservations; failures don't replenish budget
- Interfaces: memoryimportv1, workspace-qualified HTTP endpoints, and offline memory import codex --dry-run --json; main checkouts and linked worktrees share the imported project layer while normal memory isolation is unchanged
More from coding & agent
- Index Network open-sources Social Discovery Protocol: agents negotiate fit on your behalf — carsonfarmer · 2026-10-23
- Installed a formal-verification skill: it silently installed Java, MacBook runs 10 degrees hotter — zealcaiden · 2026-10-23
- OpenAI's Decisions API hits GA: near real-time routing up to 10x faster than GPT-6 Luna — OpenAIDevs · 2026-10-23
- Ruff author warns: upgrade Node/Bun now — --use-system-ca still accepts the eMudhra cert — charliermarsh · 2026-10-23
- Proof of Agent: paste one prompt and your AI mines, grades, and self-governs — adamamcbride · 2026-10-23
- DeepLearning.AI weekly: GPT-6.1 hits 97.9% byte parity on COBOL migration — DeepLearningAI · 2026-10-23