MemoryOS 2.0.27__tar.gz → 2.0.32__tar.gz

{memoryos-2.0.27 → memoryos-2.0.32}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: MemoryOS
-Version: 2.0.27
+Version: 2.0.32
 Summary: Intelligence Begins with Memory
 License: Apache-2.0
 License-File: LICENSE
@@ -30,7 +30,7 @@ Provides-Extra: tavily
 Provides-Extra: tree-mem
 Requires-Dist: alibabacloud-oss-v2 (>=1.2.2,<1.2.3) ; extra == "all"
 Requires-Dist: alibabacloud-oss-v2 (>=1.2.2,<1.2.3) ; extra == "skill-mem"
-Requires-Dist: cachetools (>=6.0.0) ; extra == "all"
+Requires-Dist: cachetools (>=6.0.0)
 Requires-Dist: chonkie (>=1.0.7,<2.0.0) ; extra == "all"
 Requires-Dist: chonkie (>=1.0.7,<2.0.0) ; extra == "mem-reader"
 Requires-Dist: concurrent-log-handler (>=0.9.28,<1.0.0)
@@ -110,9 +110,16 @@ Description-Content-Type: text/markdown
 
 
 <div align="center">
-  <img width="1660" height="664" alt="MemOS Plugin Banner" src="https://github.com/user-attachments/assets/9d15dde2-196e-4f71-a364-dd5a33062117" />
+  <img width="1660" alt="MemOS agent ecosystem: OpenClaw, Hermes, and DeepSeek Harness" src="./assets/readme/memos-agent-ecosystem.png" />
 </div>
 
+> [!TIP]
+> **New: Connect MemOS to DeepSeek Harness (`dsh`)**
+>
+> Add automatic recall, background capture, hybrid retrieval, and a local Memory Viewer to DeepSeek Harness—powered by the same MemOS core used across agent ecosystems.
+>
+> **[Get started →](#memos-plugin)**
+
 ---
 
 ## 👾 MemOS: Memory Operating System for LLM & AI Agents
@@ -130,6 +137,9 @@ Description-Content-Type: text/markdown
 
 ### News
 
+- **2026-08-17** · 🐋 **MemOS Connects with DeepSeek Harness**
+  MemOS now brings persistent memory to **DeepSeek Harness** through both local and cloud plugins. DSH can automatically recall relevant context before a task and retain new experience after a successful turn, without modifying its core.
+
 - **2026-07-02** · 🏆 **MemOS Advances Agent and User Memory Benchmarks**
   With MemOS, **OpenClaw** improves average task completion from **36.63% to 50.87%** across five agent tasks. MemOS also achieves **88.83 on LoCoMo** and **89.20 on LongMemEval**, and leads in **OmniMemEval**, a unified evaluation of 14 commercial memory products across ten datasets.
 
@@ -138,7 +148,7 @@ Description-Content-Type: text/markdown
 
 - **2026-04-10** · 👧🏻 **MemOS Hermes Agent Local Plugin**
   Official Hermes Agent memory plugins launched: Hybrid retrieval (FTS5 + vector), smart dedup, tiered skill evolution, multi-agent collaboration. 100% local, zero cloud dependency.
-  
+
 - **2026-03-08** · 🦞 **MemOS OpenClaw Plugin — Cloud & Local**
   Official OpenClaw memory plugins launched. **Cloud Plugin**: hosted memory service with 72% lower token usage and multi-agent memory sharing ([MemOS-Cloud-OpenClaw-Plugin](https://github.com/MemTensor/MemOS-Cloud-OpenClaw-Plugin)). **Local Plugin** (`v1.0.0`): 100% on-device memory with persistent SQLite, hybrid search (FTS5 + vector), task summarization & skill evolution, multi-agent collaboration, and a full Memory Viewer dashboard.
 
@@ -177,12 +187,12 @@ MemOS gives AI agents long-term memory. Common uses:
 MemOS is built around four entry points. Pick the one that matches your scenario.
 
 
-|              | Cloud API               | Self-Host          | OpenClaw Cloud Plugin    | Local Plugin                    |
-| ------------ | ----------------------- | ------------------ | ------------------------ | ------------------------------- |
-| Best for     | Your app, fully managed | Teams on own infra | OpenClaw users, zero ops | Hermes/OpenClaw, 100% on-device |
-| Setup        | Get an API key          | docker compose up  | openclaw plugins install | npm install + config            |
-| Infra needed | None (hosted)           | Neo4j + Qdrant     | None (uses MemOS Cloud)  | None (local SQLite)             |
-| Data lives   | MemOS Cloud             | Your servers       | MemOS Cloud              | Your machine                    |
+|              | Cloud API               | Self-Host          | MemOS Cloud Plugin       | Local Plugin                                     |
+| ------------ | ----------------------- | ------------------ | ------------------------ | ------------------------------------------------ |
+| Best for     | Your app, fully managed | Teams on own infra | OpenClaw users, zero ops | DeepSeek Harness, Hermes, or OpenClaw; on-device |
+| Setup        | Get an API key          | docker compose up  | openclaw plugins install | npm install + agent-specific setup               |
+| Infra needed | None (hosted)           | Neo4j + Qdrant     | None (uses MemOS Cloud)  | None (local SQLite)                              |
+| Data lives   | MemOS Cloud             | Your servers       | MemOS Cloud              | Your machine                                     |
 
 ### ☁️ Use the Cloud API (Hosted)
 
@@ -249,7 +259,7 @@ cd src
 uvicorn memos.api.server_api:app --host 0.0.0.0 --port 8000 --workers 1
 ```
 
-See `[docker/.env.example](./docker/.env.example)` for all configuration options (LLM provider, embedder, vector DB, graph DB, scheduler). The full deployment guide is at [https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/](https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/).
+See [`docker/.env.example-full`](./docker/.env.example-full) for all configuration options (LLM provider, embedder, vector DB, graph DB, scheduler). The full deployment guide is at [https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/](https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/).
 
 **Try the API:**
 
@@ -283,18 +293,20 @@ res = requests.post(f"{base}/search", headers=headers, data=json.dumps({
 print(res.json())
 ```
 
+<a id="memos-plugin"></a>
+
 ### 🧠 MemOS Plugin: Persistent Memory for Your AI Agents ✨
 
-Your OpenClaw and Hermes Agents now have **the best** memory system — choose ***Cloud Service*** or ***Self-hosted*** to get started 🏃🏻
+MemOS gives OpenClaw, Hermes, and DeepSeek Harness a shared local memory core; the managed ***MemOS Cloud Plugin*** is available for OpenClaw and DeepSeek Harness 🏃🏻
 
 | 🔌 Plugin                                                                                                     | 💡 Core Features | 🧩 Resources                                                                                                                                                                                                                                                                   |
 | ------------------------------------------------------------------------------------------------------------- | ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
 | 🧠 **[memos-local-plugin 2.0](https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin)**         |                  | 🌐 [Website](https://memos-claw.openmem.net/) · 📖 [Docs](https://memos-docs.openmem.net/cn/openclaw/local_plugin) · 🐙 [GitHub](https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin) · 📦 [NPM](https://www.npmjs.com/package/@memtensor/memos-local-plugin) |
-| ☁️ **[OpenClaw Cloud Plugin](https://github.com/MemTensor/MemOS/tree/main/apps/MemOS-Cloud-OpenClaw-Plugin)** |                  | 🖥️ [MemOS Dashboard](https://memos-dashboard.openmem.net/login/) · 📖 [Full Tutorial](https://memos-docs.openmem.net/openclaw/guide#_4-update-plugin)                                                                                                                         |
+| ☁️ **[MemOS Cloud Plugin](https://github.com/MemTensor/MemOS/tree/main/apps/MemOS-Cloud-OpenClaw-Plugin)**    |                  | 🖥️ [MemOS Dashboard](https://memos-dashboard.openmem.net/login/) · 📖 [Full Tutorial](https://memos-docs.openmem.net/openclaw/guide#_4-update-plugin)                                                                                                                         |
 
-#### 1. OpenClaw Cloud Plugin
+#### 1. MemOS Cloud Plugin
 
-You use OpenClaw and want persistent memory via MemOS Cloud — no infrastructure to run.
+Use MemOS Cloud for persistent memory in OpenClaw or DeepSeek Harness — no infrastructure to run.
 
 - **Repo:** [MemTensor/MemOS ·](https://github.com/MemTensor/MemOS/tree/main/apps/MemOS-Cloud-OpenClaw-Plugin) `apps/MemOS-Cloud-OpenClaw-Plugin`
 - **NPM:** `[@memtensor/memos-cloud-openclaw-plugin](https://www.npmjs.com/package/@memtensor/memos-cloud-openclaw-plugin)`
@@ -310,16 +322,55 @@ openclaw gateway restart
 
 The plugin recalls memories from MemOS Cloud before each agent run and saves new messages back after the run ends.
 
-#### 2. Local Plugin (memos-local-plugin 2.0)
+##### DeepSeek Harness
+
+Connect DeepSeek Harness to MemOS Cloud through its native plugin mechanism. Before the first model step of each user request, the plugin recalls relevant cloud memories; after a successful turn, it saves the new user and assistant messages back to MemOS Cloud.
+
+1. [Create a MemOS API Key](https://memos-dashboard.openmem.net/cn/apikeys/).
+2. Install the cloud plugin into the default DSH `web` profile:
+
+   ```bash
+   npx @deepseek-ai/dsh plugin --profile web add @memtensor/memos-cloud-dsh-plugin@latest
+   ```
+
+3. Add the API Key to `~/.dsh/.credentials.yaml`:
+
+   ```yaml
+   MEMOS_API_KEY: mpg-your-key
+   ```
+
+4. Add the minimal plugin configuration to `~/.dsh/settings.yaml`:
 
-You use Hermes Agent or OpenClaw and want 100% on-device memory — nothing leaves your machine.
+   ```yaml
+   memos-cloud:
+     apiKeyEnv: MEMOS_API_KEY
+   ```
+
+5. Restart the DSH Web profile:
+
+   ```bash
+   npx @deepseek-ai/dsh web
+   ```
+
+The cloud plugin is fail-open: a temporary MemOS Cloud outage does not interrupt the current DSH task.
+
+#### 2. Local Plugin (OpenClaw, Hermes, and DeepSeek Harness)
+
+You use DeepSeek Harness, Hermes Agent, or OpenClaw and want 100% on-device memory — nothing leaves your machine.
 
 - **Repo:** [MemTensor/MemOS ·](https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin) `apps/memos-local-plugin`
 - **NPM:** `[@memtensor/memos-local-plugin](https://www.npmjs.com/package/@memtensor/memos-local-plugin)`
 - **Docs:** [https://memos-docs.openmem.net/cn/openclaw/local_plugin](https://memos-docs.openmem.net/cn/openclaw/local_plugin)
+- **DeepSeek Harness:** [integration details](./apps/memos-local-plugin/adapters/deepseek-harness/README.md)
 - **Viewer dashboard:** see `apps/memos-local-plugin/viewer/`
 
-Install (macOS / Linux):
+Install for DeepSeek Harness (macOS / Linux):
+
+```bash
+curl -fsSL https://raw.githubusercontent.com/MemTensor/MemOS/main/apps/memos-local-plugin/install.sh | bash -s -- --agent dsh --profile web
+```
+
+Install for OpenClaw or Hermes (macOS / Linux):
 
 ```bash
 curl -fsSL https://raw.githubusercontent.com/MemTensor/MemOS/main/apps/memos-local-plugin/install.sh | bash
@@ -331,7 +382,7 @@ Install (Windows PowerShell):
 irm https://raw.githubusercontent.com/MemTensor/MemOS/main/apps/memos-local-plugin/install.ps1 -OutFile "$env:TEMP\memos-install.ps1"; powershell -ExecutionPolicy Bypass -File "$env:TEMP\memos-install.ps1"
 ```
 
-Requires Node.js and an already-installed OpenClaw or Hermes. The installer auto-detects OpenClaw and Hermes, deploys the plugin to the right agent home (`~/.hermes/plugins/` or `~/.openclaw/plugins/`), writes the initial `config.yaml`, and restarts the agent runtime.
+Requires Node.js and an already-installed DeepSeek Harness, OpenClaw, or Hermes. The installer deploys MemOS to the selected agent runtime; the DeepSeek Harness target installs it as an out-of-tree DSH bundle, while the OpenClaw and Hermes targets write the initial `config.yaml` in their respective agent homes.
 
 Features: hybrid retrieval (FTS5 + vector), smart dedup, tiered skill evolution (L1 traces / L2 policies / L3 world model), multi-agent collaboration, local-first SQLite storage.
 

{memoryos-2.0.27 → memoryos-2.0.32}/README.md

@@ -26,9 +26,16 @@
 
 
 <div align="center">
-  <img width="1660" height="664" alt="MemOS Plugin Banner" src="https://github.com/user-attachments/assets/9d15dde2-196e-4f71-a364-dd5a33062117" />
+  <img width="1660" alt="MemOS agent ecosystem: OpenClaw, Hermes, and DeepSeek Harness" src="./assets/readme/memos-agent-ecosystem.png" />
 </div>
 
+> [!TIP]
+> **New: Connect MemOS to DeepSeek Harness (`dsh`)**
+>
+> Add automatic recall, background capture, hybrid retrieval, and a local Memory Viewer to DeepSeek Harness—powered by the same MemOS core used across agent ecosystems.
+>
+> **[Get started →](#memos-plugin)**
+
 ---
 
 ## 👾 MemOS: Memory Operating System for LLM & AI Agents
@@ -46,6 +53,9 @@
 
 ### News
 
+- **2026-08-17** · 🐋 **MemOS Connects with DeepSeek Harness**
+  MemOS now brings persistent memory to **DeepSeek Harness** through both local and cloud plugins. DSH can automatically recall relevant context before a task and retain new experience after a successful turn, without modifying its core.
+
 - **2026-07-02** · 🏆 **MemOS Advances Agent and User Memory Benchmarks**
   With MemOS, **OpenClaw** improves average task completion from **36.63% to 50.87%** across five agent tasks. MemOS also achieves **88.83 on LoCoMo** and **89.20 on LongMemEval**, and leads in **OmniMemEval**, a unified evaluation of 14 commercial memory products across ten datasets.
 
@@ -54,7 +64,7 @@
 
 - **2026-04-10** · 👧🏻 **MemOS Hermes Agent Local Plugin**
   Official Hermes Agent memory plugins launched: Hybrid retrieval (FTS5 + vector), smart dedup, tiered skill evolution, multi-agent collaboration. 100% local, zero cloud dependency.
-  
+
 - **2026-03-08** · 🦞 **MemOS OpenClaw Plugin — Cloud & Local**
   Official OpenClaw memory plugins launched. **Cloud Plugin**: hosted memory service with 72% lower token usage and multi-agent memory sharing ([MemOS-Cloud-OpenClaw-Plugin](https://github.com/MemTensor/MemOS-Cloud-OpenClaw-Plugin)). **Local Plugin** (`v1.0.0`): 100% on-device memory with persistent SQLite, hybrid search (FTS5 + vector), task summarization & skill evolution, multi-agent collaboration, and a full Memory Viewer dashboard.
 
@@ -93,12 +103,12 @@ MemOS gives AI agents long-term memory. Common uses:
 MemOS is built around four entry points. Pick the one that matches your scenario.
 
 
-|              | Cloud API               | Self-Host          | OpenClaw Cloud Plugin    | Local Plugin                    |
-| ------------ | ----------------------- | ------------------ | ------------------------ | ------------------------------- |
-| Best for     | Your app, fully managed | Teams on own infra | OpenClaw users, zero ops | Hermes/OpenClaw, 100% on-device |
-| Setup        | Get an API key          | docker compose up  | openclaw plugins install | npm install + config            |
-| Infra needed | None (hosted)           | Neo4j + Qdrant     | None (uses MemOS Cloud)  | None (local SQLite)             |
-| Data lives   | MemOS Cloud             | Your servers       | MemOS Cloud              | Your machine                    |
+|              | Cloud API               | Self-Host          | MemOS Cloud Plugin       | Local Plugin                                     |
+| ------------ | ----------------------- | ------------------ | ------------------------ | ------------------------------------------------ |
+| Best for     | Your app, fully managed | Teams on own infra | OpenClaw users, zero ops | DeepSeek Harness, Hermes, or OpenClaw; on-device |
+| Setup        | Get an API key          | docker compose up  | openclaw plugins install | npm install + agent-specific setup               |
+| Infra needed | None (hosted)           | Neo4j + Qdrant     | None (uses MemOS Cloud)  | None (local SQLite)                              |
+| Data lives   | MemOS Cloud             | Your servers       | MemOS Cloud              | Your machine                                     |
 
 ### ☁️ Use the Cloud API (Hosted)
 
@@ -165,7 +175,7 @@ cd src
 uvicorn memos.api.server_api:app --host 0.0.0.0 --port 8000 --workers 1
 ```
 
-See `[docker/.env.example](./docker/.env.example)` for all configuration options (LLM provider, embedder, vector DB, graph DB, scheduler). The full deployment guide is at [https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/](https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/).
+See [`docker/.env.example-full`](./docker/.env.example-full) for all configuration options (LLM provider, embedder, vector DB, graph DB, scheduler). The full deployment guide is at [https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/](https://memos-docs.openmem.net/open_source/getting_started/rest_api_server/).
 
 **Try the API:**
 
@@ -199,18 +209,20 @@ res = requests.post(f"{base}/search", headers=headers, data=json.dumps({
 print(res.json())
 ```
 
+<a id="memos-plugin"></a>
+
 ### 🧠 MemOS Plugin: Persistent Memory for Your AI Agents ✨
 
-Your OpenClaw and Hermes Agents now have **the best** memory system — choose ***Cloud Service*** or ***Self-hosted*** to get started 🏃🏻
+MemOS gives OpenClaw, Hermes, and DeepSeek Harness a shared local memory core; the managed ***MemOS Cloud Plugin*** is available for OpenClaw and DeepSeek Harness 🏃🏻
 
 | 🔌 Plugin                                                                                                     | 💡 Core Features | 🧩 Resources                                                                                                                                                                                                                                                                   |
 | ------------------------------------------------------------------------------------------------------------- | ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
 | 🧠 **[memos-local-plugin 2.0](https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin)**         |                  | 🌐 [Website](https://memos-claw.openmem.net/) · 📖 [Docs](https://memos-docs.openmem.net/cn/openclaw/local_plugin) · 🐙 [GitHub](https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin) · 📦 [NPM](https://www.npmjs.com/package/@memtensor/memos-local-plugin) |
-| ☁️ **[OpenClaw Cloud Plugin](https://github.com/MemTensor/MemOS/tree/main/apps/MemOS-Cloud-OpenClaw-Plugin)** |                  | 🖥️ [MemOS Dashboard](https://memos-dashboard.openmem.net/login/) · 📖 [Full Tutorial](https://memos-docs.openmem.net/openclaw/guide#_4-update-plugin)                                                                                                                         |
+| ☁️ **[MemOS Cloud Plugin](https://github.com/MemTensor/MemOS/tree/main/apps/MemOS-Cloud-OpenClaw-Plugin)**    |                  | 🖥️ [MemOS Dashboard](https://memos-dashboard.openmem.net/login/) · 📖 [Full Tutorial](https://memos-docs.openmem.net/openclaw/guide#_4-update-plugin)                                                                                                                         |
 
-#### 1. OpenClaw Cloud Plugin
+#### 1. MemOS Cloud Plugin
 
-You use OpenClaw and want persistent memory via MemOS Cloud — no infrastructure to run.
+Use MemOS Cloud for persistent memory in OpenClaw or DeepSeek Harness — no infrastructure to run.
 
 - **Repo:** [MemTensor/MemOS ·](https://github.com/MemTensor/MemOS/tree/main/apps/MemOS-Cloud-OpenClaw-Plugin) `apps/MemOS-Cloud-OpenClaw-Plugin`
 - **NPM:** `[@memtensor/memos-cloud-openclaw-plugin](https://www.npmjs.com/package/@memtensor/memos-cloud-openclaw-plugin)`
@@ -226,16 +238,55 @@ openclaw gateway restart
 
 The plugin recalls memories from MemOS Cloud before each agent run and saves new messages back after the run ends.
 
-#### 2. Local Plugin (memos-local-plugin 2.0)
+##### DeepSeek Harness
+
+Connect DeepSeek Harness to MemOS Cloud through its native plugin mechanism. Before the first model step of each user request, the plugin recalls relevant cloud memories; after a successful turn, it saves the new user and assistant messages back to MemOS Cloud.
+
+1. [Create a MemOS API Key](https://memos-dashboard.openmem.net/cn/apikeys/).
+2. Install the cloud plugin into the default DSH `web` profile:
+
+   ```bash
+   npx @deepseek-ai/dsh plugin --profile web add @memtensor/memos-cloud-dsh-plugin@latest
+   ```
+
+3. Add the API Key to `~/.dsh/.credentials.yaml`:
+
+   ```yaml
+   MEMOS_API_KEY: mpg-your-key
+   ```
+
+4. Add the minimal plugin configuration to `~/.dsh/settings.yaml`:
 
-You use Hermes Agent or OpenClaw and want 100% on-device memory — nothing leaves your machine.
+   ```yaml
+   memos-cloud:
+     apiKeyEnv: MEMOS_API_KEY
+   ```
+
+5. Restart the DSH Web profile:
+
+   ```bash
+   npx @deepseek-ai/dsh web
+   ```
+
+The cloud plugin is fail-open: a temporary MemOS Cloud outage does not interrupt the current DSH task.
+
+#### 2. Local Plugin (OpenClaw, Hermes, and DeepSeek Harness)
+
+You use DeepSeek Harness, Hermes Agent, or OpenClaw and want 100% on-device memory — nothing leaves your machine.
 
 - **Repo:** [MemTensor/MemOS ·](https://github.com/MemTensor/MemOS/tree/main/apps/memos-local-plugin) `apps/memos-local-plugin`
 - **NPM:** `[@memtensor/memos-local-plugin](https://www.npmjs.com/package/@memtensor/memos-local-plugin)`
 - **Docs:** [https://memos-docs.openmem.net/cn/openclaw/local_plugin](https://memos-docs.openmem.net/cn/openclaw/local_plugin)
+- **DeepSeek Harness:** [integration details](./apps/memos-local-plugin/adapters/deepseek-harness/README.md)
 - **Viewer dashboard:** see `apps/memos-local-plugin/viewer/`
 
-Install (macOS / Linux):
+Install for DeepSeek Harness (macOS / Linux):
+
+```bash
+curl -fsSL https://raw.githubusercontent.com/MemTensor/MemOS/main/apps/memos-local-plugin/install.sh | bash -s -- --agent dsh --profile web
+```
+
+Install for OpenClaw or Hermes (macOS / Linux):
 
 ```bash
 curl -fsSL https://raw.githubusercontent.com/MemTensor/MemOS/main/apps/memos-local-plugin/install.sh | bash
@@ -247,7 +298,7 @@ Install (Windows PowerShell):
 irm https://raw.githubusercontent.com/MemTensor/MemOS/main/apps/memos-local-plugin/install.ps1 -OutFile "$env:TEMP\memos-install.ps1"; powershell -ExecutionPolicy Bypass -File "$env:TEMP\memos-install.ps1"
 ```
 
-Requires Node.js and an already-installed OpenClaw or Hermes. The installer auto-detects OpenClaw and Hermes, deploys the plugin to the right agent home (`~/.hermes/plugins/` or `~/.openclaw/plugins/`), writes the initial `config.yaml`, and restarts the agent runtime.
+Requires Node.js and an already-installed DeepSeek Harness, OpenClaw, or Hermes. The installer deploys MemOS to the selected agent runtime; the DeepSeek Harness target installs it as an out-of-tree DSH bundle, while the OpenClaw and Hermes targets write the initial `config.yaml` in their respective agent homes.
 
 Features: hybrid retrieval (FTS5 + vector), smart dedup, tiered skill evolution (L1 traces / L2 policies / L3 world model), multi-agent collaboration, local-first SQLite storage.
 

{memoryos-2.0.27 → memoryos-2.0.32}/pyproject.toml

@@ -4,7 +4,7 @@
 ##############################################################################
 
 name = "MemoryOS"
-version = "2.0.27"
+version = "2.0.32"
 description = "Intelligence Begins with Memory"
 license = {text = "Apache-2.0"}
 readme = "README.md"
@@ -48,6 +48,7 @@ dependencies = [
     "python-dateutil (>=2.9.0.post0,<3.0.0)",
     "prometheus-client (>=0.23.1,<0.24.0)",
     "concurrent-log-handler (>=0.9.28,<1.0.0)",  # Process-safe rotating file handler
+    "cachetools (>=6.0.0)",  # Embedding request caching
 ]
 
 [project.urls]
@@ -128,7 +129,6 @@ all = [
     "datasketch (>=1.6.5,<2.0.0)",
     "jieba (>=0.38.1,<0.42.1)",
     "rank-bm25 (>=0.2.2)",
-    "cachetools (>=6.0.0)",
     # NOT exist in the above optional groups
     # Because they are either huge-size dependencies or infrequently used dependencies.
     # We kindof don't want users to install them.

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "2.0.27"
+__version__ = "2.0.32"
 
 from memos.configs.mem_cube import GeneralMemCubeConfig
 from memos.configs.mem_os import MOSConfig

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/api/config.py

@@ -463,6 +463,28 @@ class APIConfig:
         # Fallback to memreader config (same behavior as before for users who don't customize)
         return APIConfig.get_memreader_config()
 
+    @staticmethod
+    def get_suggestion_llm_config() -> dict[str, Any]:
+        """Get LLM configuration for suggestion query generation.
+
+        Fallback chain: SUGGESTION_MODEL -> MEMREADER_GENERAL_MODEL -> memreader config.
+        """
+        suggestion_model = os.getenv("SUGGESTION_MODEL")
+        if suggestion_model:
+            return APIConfig._build_provider_llm_config(suggestion_model)
+        return APIConfig.get_memreader_general_llm_config()
+
+    @staticmethod
+    def get_scheduler_llm_config() -> dict[str, Any]:
+        """Get LLM configuration for scheduler-owned LLM tasks.
+
+        Fallback chain: MEMSCHEDULER_MODEL -> MEMREADER_GENERAL_MODEL.
+        """
+        scheduler_model = os.getenv("MEMSCHEDULER_MODEL")
+        if scheduler_model:
+            return APIConfig._build_provider_llm_config(scheduler_model)
+        return APIConfig.get_memreader_general_llm_config()
+
     @staticmethod
     def get_image_parser_llm_config() -> dict[str, Any]:
         """Get LLM configuration for image parsing (requires vision model).
@@ -622,6 +644,7 @@ class APIConfig:
                     "provider": os.getenv("MOS_EMBEDDER_PROVIDER", "openai"),
                     "api_key": os.getenv("MOS_EMBEDDER_API_KEY", "sk-xxxx"),
                     "model_name_or_path": os.getenv("MOS_EMBEDDER_MODEL", "text-embedding-3-large"),
+                    "embedding_dims": int(os.getenv("EMBEDDING_DIMENSION", "1024")),
                     "headers_extra": json.loads(os.getenv("MOS_EMBEDDER_HEADERS_EXTRA", "{}")),
                     "base_url": os.getenv("MOS_EMBEDDER_API_BASE", "http://openai.com"),
                     "backup_client": os.getenv("MOS_EMBEDDER_BACKUP_CLIENT", "false").lower()

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/api/handlers/chat_handler.py

@@ -92,6 +92,7 @@ class ChatHandler(BaseHandler):
 
         self.chat_llms = chat_llms
         self.playground_chat_llms = playground_chat_llms or chat_llms
+        self.suggestion_llm = getattr(dependencies, "suggestion_llm", dependencies.llm)
         self.search_handler = search_handler
         self.add_handler = add_handler
         self.online_bot = online_bot
@@ -1232,7 +1233,7 @@ class ChatHandler(BaseHandler):
                 },
                 {"role": "user", "content": further_suggestion_prompt},
             ]
-            response = self.llm.generate(message_list)
+            response = self.suggestion_llm.generate(message_list)
             clean_response = clean_json_response(response)
             response_json = json.loads(clean_response)
             return response_json["query"]

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/api/handlers/component_init.py

@@ -21,6 +21,8 @@ from memos.api.handlers.config_builders import (
     build_mem_reader_config,
     build_nli_client_config,
     build_reranker_config,
+    build_scheduler_llm_config,
+    build_suggestion_llm_config,
 )
 from memos.configs.mem_scheduler import SchedulerConfigFactory
 from memos.embedders.factory import EmbedderFactory
@@ -158,6 +160,8 @@ def init_server() -> dict[str, Any]:
     graph_db_config = build_graph_db_config()
     llm_config = build_llm_config()
     feedback_llm_config = build_feedback_llm_config()
+    suggestion_llm_config = build_suggestion_llm_config()
+    scheduler_llm_config = build_scheduler_llm_config()
     chat_llm_config = build_chat_llm_config()
     playground_chat_llm_config = build_chat_llm_config("PLAYGROUND_CHAT_MODEL_LIST")
     embedder_config = build_embedder_config()
@@ -173,6 +177,8 @@ def init_server() -> dict[str, Any]:
     graph_db = GraphStoreFactory.from_config(graph_db_config)
     llm = LLMFactory.from_config(llm_config)
     feedback_llm = LLMFactory.from_config(feedback_llm_config)
+    suggestion_llm = LLMFactory.from_config(suggestion_llm_config)
+    scheduler_llm = LLMFactory.from_config(scheduler_llm_config)
     chat_llms = (
         _init_chat_llms(chat_llm_config)
         if os.getenv("ENABLE_CHAT_API", "false") == "true"
@@ -281,7 +287,7 @@ def init_server() -> dict[str, Any]:
     )
     mem_scheduler: OptimizedScheduler = SchedulerFactory.from_config(scheduler_config)
     mem_scheduler.initialize_modules(
-        chat_llm=llm,
+        chat_llm=scheduler_llm,
         process_llm=mem_reader.general_llm,
         db_engine=BaseDBManager.create_default_sqlite_engine(),
         mem_reader=mem_reader,
@@ -325,6 +331,8 @@ def init_server() -> dict[str, Any]:
         "graph_db": graph_db,
         "mem_reader": mem_reader,
         "llm": llm,
+        "suggestion_llm": suggestion_llm,
+        "scheduler_llm": scheduler_llm,
         "chat_llms": chat_llms,
         "playground_chat_llms": playground_chat_llms,
         "embedder": embedder,

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/api/handlers/config_builders.py

@@ -167,6 +167,26 @@ def build_feedback_llm_config() -> dict[str, Any]:
     return LLMConfigFactory.model_validate(APIConfig.get_feedback_llm_config())
 
 
+def build_suggestion_llm_config() -> dict[str, Any]:
+    """
+    Build suggestion LLM configuration.
+
+    Returns:
+        Validated suggestion LLM configuration dictionary
+    """
+    return LLMConfigFactory.model_validate(APIConfig.get_suggestion_llm_config())
+
+
+def build_scheduler_llm_config() -> dict[str, Any]:
+    """
+    Build scheduler LLM configuration.
+
+    Returns:
+        Validated scheduler LLM configuration dictionary
+    """
+    return LLMConfigFactory.model_validate(APIConfig.get_scheduler_llm_config())
+
+
 def build_internet_retriever_config() -> dict[str, Any]:
     """
     Build internet retriever configuration.

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/api/routers/server_router.py

@@ -97,6 +97,7 @@ cube_handler = CubeHandler(dependencies)
 # (These can be accessed from the components dict without unpacking all of them)
 mem_scheduler: BaseScheduler = components["mem_scheduler"]
 llm = components["llm"]
+suggestion_llm = components.get("suggestion_llm", llm)
 naive_mem_cube = components["naive_mem_cube"]
 redis_client = components["redis_client"]
 status_tracker = TaskStatusTracker(redis_client=redis_client)
@@ -327,7 +328,7 @@ def get_suggestion_queries(suggestion_req: SuggestionRequest):
         user_id=suggestion_req.mem_cube_id,
         language=suggestion_req.language,
         message=suggestion_req.message,
-        llm=llm,
+        llm=suggestion_llm,
         naive_mem_cube=naive_mem_cube,
     )
 

{memoryos-2.0.27 → memoryos-2.0.32}/src/memos/chunkers/base.py

@@ -1,4 +1,5 @@
 import re
+import uuid
 
 from abc import ABC, abstractmethod
 
@@ -17,6 +18,12 @@ class Chunk:
 class BaseChunker(ABC):
     """Base class for all text chunkers."""
 
+    _URL_PATTERN = r'https?://[^\s<>"{}|\\^`\[\]]+'
+    _URL_PLACEHOLDER_PREFIX = "__URL_"
+    _URL_PLACEHOLDER_PATTERN = re.compile(
+        rf"{re.escape(_URL_PLACEHOLDER_PREFIX)}(?:[0-9a-f]+_)?\d+__"
+    )
+
     @abstractmethod
     def __init__(self, config: BaseChunkerConfig):
         """Initialize the chunker with the given configuration."""
@@ -25,6 +32,17 @@ class BaseChunker(ABC):
     def chunk(self, text: str) -> list[Chunk]:
         """Chunk the given text into smaller chunks."""
 
+    @classmethod
+    def _new_url_placeholder_prefix(cls, text: str) -> str:
+        """Return a placeholder namespace that cannot collide with the input."""
+        if cls._URL_PLACEHOLDER_PREFIX not in text:
+            return cls._URL_PLACEHOLDER_PREFIX
+
+        while True:
+            placeholder_prefix = f"{cls._URL_PLACEHOLDER_PREFIX}{uuid.uuid4().hex[:8]}_"
+            if placeholder_prefix not in text:
+                return placeholder_prefix
+
     def protect_urls(self, text: str) -> tuple[str, dict[str, str]]:
         """
         Protect URLs in text from being split during chunking.
@@ -35,16 +53,16 @@ class BaseChunker(ABC):
         Returns:
             tuple: (Text with URLs replaced by placeholders, URL mapping dictionary)
         """
-        url_pattern = r'https?://[^\s<>"{}|\\^`\[\]]+'
-        url_map = {}
+        url_map: dict[str, str] = {}
+        placeholder_prefix = self._new_url_placeholder_prefix(text)
 
         def replace_url(match):
             url = match.group(0)
-            placeholder = f"__URL_{len(url_map)}__"
+            placeholder = f"{placeholder_prefix}{len(url_map)}__"
             url_map[placeholder] = url
             return placeholder
 
-        protected_text = re.sub(url_pattern, replace_url, text)
+        protected_text = re.sub(self._URL_PATTERN, replace_url, text)
         return protected_text, url_map
 
     def restore_urls(self, text: str, url_map: dict[str, str]) -> str:
@@ -58,8 +76,6 @@ class BaseChunker(ABC):
         Returns:
             str: Text with URLs restored
         """
-        restored_text = text
-        for placeholder, url in url_map.items():
-            restored_text = restored_text.replace(placeholder, url)
-
-        return restored_text
+        return self._URL_PLACEHOLDER_PATTERN.sub(
+            lambda match: url_map.get(match.group(0), match.group(0)), text
+        )