@meshagent/meshagent 0.39.5 → 0.39.6

package/CHANGELOG.md

@@ -1,3 +1,9 @@
+## [0.39.6]
+- Chat thread message sending now supports an “agent messages” mode: it selects participants that advertise agent-message support and sends `agent-message` payloads using `meshagent.agent.turn.start` / `meshagent.agent.turn.steer` types (including turn/thread scoping), with Promise-based sending and cancellation when recipients never materialize.
+- Chat UI/logic now determines the correct outbound message type (chat vs steer) and turn context from thread status, and passes that into message sending.
+- DatasetsClient now adds strongly typed `importFromStorage` and `exportToStorage` APIs with dataset storage format + import mode options, optional `namespace`/`branch` scoping, and `batch_size` support that is omitted when unset.
+- Unit tests were updated to verify the new import/export request payload shapes and defaults.
+
 ## [0.39.5]
 - Stability
 

package/README.md

@@ -6,7 +6,6 @@ MeshAgent is your platform to create, deploy, and manage AI agents collaborative
 
 MeshAgent removes the infrastructure headaches of building and shipping AI Agents. It spins up secure, real-time "Rooms" that connect humans, agents, and shared context -- letting you launch, share, and refine agents in hours instead of weeks.
 
-
 ---
 
 **Documentation**: [docs.meshagent.com](https://docs.meshagent.com/)
@@ -23,6 +22,31 @@ Install the MeshAgent CLI, connect it to your account, then run a minimal toolki
 
 ### 1. Install the MeshAgent CLI
 
+If you are using a Python-based install path, MeshAgent currently requires **Python 3.13**. Python 3.14 is not supported yet, and earlier versions are not tested.
+
+If you do not already have Python 3.13, the simplest setup is:
+
+#### Install uv on Linux or macOS
+```bash
+curl -LsSf https://astral.sh/uv/install.sh | sh
+```
+
+#### or Install uv on windows (PowerShell):
+```bash
+powershelliwr -useb https://astral.sh/uv/install.ps1 | iex
+```
+
+
+#### Download and manage Python 3.13
+```bash
+uv python install 3.13
+```
+
+#### Verify the version
+```bash
+uv run --python 3.13 python --version
+```
+
 Use the install path that matches your environment:
 
 ```bash
@@ -34,8 +58,7 @@ brew install meshagent
 choco install meshagent
 
 # Other platforms
-pipx install "meshagent[cli]" --include-deps
-pipx ensurepath
+uv tool install --python 3.13 "meshagent[cli]"
 ```
 
 Then sign in and activate a project:
@@ -144,11 +167,68 @@ main().catch((error) => {
 ### 4. Run the example
 
 ```bash
-meshagent room connect --project-id <project-id> --room <room-name> -- node example.js
+meshagent room connect --room <your-room-name> -- node example.js
 ```
 
 `meshagent room connect` starts the local Node process with the MeshAgent room environment already configured, so `RoomClient()` can connect without any extra setup in your code. If you already activated a project with `meshagent setup`, you can omit `--project-id`.
 
+## Deploy With a Dockerfile
+
+For deployment, package your toolkit as a long-running Node process. Use the same `RoomClient` and `startHostedToolkit` setup from the local example, but do not invoke the tool and exit; keep the process alive until it receives `SIGTERM` or `SIGINT`.
+
+Create a `Dockerfile` for your toolkit service:
+
+```dockerfile
+FROM node:22-slim
+
+WORKDIR /app
+COPY package*.json ./
+RUN npm ci --omit=dev
+COPY . .
+
+CMD ["node", "service.js"]
+```
+
+Build and push the image to a registry that MeshAgent can pull:
+
+```bash
+docker buildx build . \
+  -t "<REGISTRY>/<NAMESPACE>/simple-echo-toolkit:<TAG>" \
+  --platform linux/amd64 \
+  --push
+```
+
+Then create a `meshagent.yaml` service manifest that runs the image and injects a room token:
+
+```yaml
+kind: Service
+version: v1
+metadata:
+  name: simple-echo-toolkit
+  description: "Simple Echo Toolkit hosted from a Node container"
+  annotations:
+    meshagent.service.id: simple-echo-toolkit
+agents:
+  - name: simple-echo-toolkit
+    description: "Hosts the Simple Echo Toolkit"
+container:
+  image: "<REGISTRY>/<NAMESPACE>/simple-echo-toolkit:<TAG>"
+  command: node service.js
+  environment:
+    - name: MESHAGENT_TOKEN
+      token:
+        identity: simple-echo-toolkit
+```
+
+Validate and deploy it to a room:
+
+```bash
+meshagent service validate --file meshagent.yaml
+meshagent service create --file meshagent.yaml --room <your-room-name>
+```
+
+After the service starts, the toolkit name you passed to `new Toolkit({ name: ... })` is available in that room.
+
 ## Next Steps and Examples
 
 To see examples of agents in action and to start building your own agents check out the MeshAgent docs at [docs.meshagent.com](https://docs.meshagent.com/)

package/dist/browser/datasets-client.d.ts

@@ -1,6 +1,8 @@
 import { Schema, Table } from "apache-arrow";
 import { RoomClient } from "./room-client";
 export type CreateMode = "create" | "overwrite" | "create_if_not_exists";
+export type DatasetStorageFormat = "auto" | "json" | "arrow" | "csv" | "tsv" | "parquet" | "excel";
+export type DatasetImportMode = "create" | "replace" | "merge";
 export interface TableRef {
     name: string;
     namespace?: string[];
@@ -222,6 +224,25 @@ export declare class DatasetsClient {
         namespace?: string[];
         branch?: string;
     }): Promise<void>;
+    importFromStorage({ table, path, mode, format, on, sheet, batchSize, namespace, branch }: {
+        table: string;
+        path: string;
+        mode?: DatasetImportMode;
+        format?: DatasetStorageFormat;
+        on?: string;
+        sheet?: string;
+        batchSize?: number;
+        namespace?: string[];
+        branch?: string;
+    }): Promise<void>;
+    exportToStorage({ table, path, format, namespace, branch, version }: {
+        table: string;
+        path: string;
+        format: DatasetStorageFormat;
+        namespace?: string[];
+        branch?: string;
+        version?: number;
+    }): Promise<void>;
     dropIndex({ table, name, namespace, branch }: {
         table: string;
         name: string;

package/dist/browser/datasets-client.js

@@ -870,6 +870,32 @@ class DatasetsClient {
             },
         });
     }
+    async importFromStorage({ table, path, mode = "create", format = "auto", on, sheet, batchSize, namespace, branch }) {
+        const input = {
+            table,
+            path,
+            mode,
+            format,
+            on: on ?? null,
+            sheet: sheet ?? null,
+            namespace: namespace ?? null,
+            branch: branch ?? null,
+        };
+        if (batchSize != null) {
+            input.batch_size = batchSize;
+        }
+        await this.invoke("import_storage", input);
+    }
+    async exportToStorage({ table, path, format, namespace, branch, version }) {
+        await this.invoke("export_storage", {
+            table,
+            path,
+            format,
+            namespace: namespace ?? null,
+            branch: branch ?? null,
+            version: version ?? null,
+        });
+    }
     async dropIndex({ table, name, namespace, branch }) {
         await this.room.invoke({
             toolkit: "dataset",

package/dist/esm/datasets-client.d.ts

@@ -1,6 +1,8 @@
 import { Schema, Table } from "apache-arrow";
 import { RoomClient } from "./room-client";
 export type CreateMode = "create" | "overwrite" | "create_if_not_exists";
+export type DatasetStorageFormat = "auto" | "json" | "arrow" | "csv" | "tsv" | "parquet" | "excel";
+export type DatasetImportMode = "create" | "replace" | "merge";
 export interface TableRef {
     name: string;
     namespace?: string[];
@@ -222,6 +224,25 @@ export declare class DatasetsClient {
         namespace?: string[];
         branch?: string;
     }): Promise<void>;
+    importFromStorage({ table, path, mode, format, on, sheet, batchSize, namespace, branch }: {
+        table: string;
+        path: string;
+        mode?: DatasetImportMode;
+        format?: DatasetStorageFormat;
+        on?: string;
+        sheet?: string;
+        batchSize?: number;
+        namespace?: string[];
+        branch?: string;
+    }): Promise<void>;
+    exportToStorage({ table, path, format, namespace, branch, version }: {
+        table: string;
+        path: string;
+        format: DatasetStorageFormat;
+        namespace?: string[];
+        branch?: string;
+        version?: number;
+    }): Promise<void>;
     dropIndex({ table, name, namespace, branch }: {
         table: string;
         name: string;

package/dist/esm/datasets-client.js

@@ -870,6 +870,32 @@ class DatasetsClient {
             },
         });
     }
+    async importFromStorage({ table, path, mode = "create", format = "auto", on, sheet, batchSize, namespace, branch }) {
+        const input = {
+            table,
+            path,
+            mode,
+            format,
+            on: on ?? null,
+            sheet: sheet ?? null,
+            namespace: namespace ?? null,
+            branch: branch ?? null,
+        };
+        if (batchSize != null) {
+            input.batch_size = batchSize;
+        }
+        await this.invoke("import_storage", input);
+    }
+    async exportToStorage({ table, path, format, namespace, branch, version }) {
+        await this.invoke("export_storage", {
+            table,
+            path,
+            format,
+            namespace: namespace ?? null,
+            branch: branch ?? null,
+            version: version ?? null,
+        });
+    }
     async dropIndex({ table, name, namespace, branch }) {
         await this.room.invoke({
             toolkit: "dataset",

package/dist/node/datasets-client.d.ts

@@ -1,6 +1,8 @@
 import { Schema, Table } from "apache-arrow";
 import { RoomClient } from "./room-client";
 export type CreateMode = "create" | "overwrite" | "create_if_not_exists";
+export type DatasetStorageFormat = "auto" | "json" | "arrow" | "csv" | "tsv" | "parquet" | "excel";
+export type DatasetImportMode = "create" | "replace" | "merge";
 export interface TableRef {
     name: string;
     namespace?: string[];
@@ -222,6 +224,25 @@ export declare class DatasetsClient {
         namespace?: string[];
         branch?: string;
     }): Promise<void>;
+    importFromStorage({ table, path, mode, format, on, sheet, batchSize, namespace, branch }: {
+        table: string;
+        path: string;
+        mode?: DatasetImportMode;
+        format?: DatasetStorageFormat;
+        on?: string;
+        sheet?: string;
+        batchSize?: number;
+        namespace?: string[];
+        branch?: string;
+    }): Promise<void>;
+    exportToStorage({ table, path, format, namespace, branch, version }: {
+        table: string;
+        path: string;
+        format: DatasetStorageFormat;
+        namespace?: string[];
+        branch?: string;
+        version?: number;
+    }): Promise<void>;
     dropIndex({ table, name, namespace, branch }: {
         table: string;
         name: string;

package/dist/node/datasets-client.js

@@ -870,6 +870,32 @@ class DatasetsClient {
             },
         });
     }
+    async importFromStorage({ table, path, mode = "create", format = "auto", on, sheet, batchSize, namespace, branch }) {
+        const input = {
+            table,
+            path,
+            mode,
+            format,
+            on: on ?? null,
+            sheet: sheet ?? null,
+            namespace: namespace ?? null,
+            branch: branch ?? null,
+        };
+        if (batchSize != null) {
+            input.batch_size = batchSize;
+        }
+        await this.invoke("import_storage", input);
+    }
+    async exportToStorage({ table, path, format, namespace, branch, version }) {
+        await this.invoke("export_storage", {
+            table,
+            path,
+            format,
+            namespace: namespace ?? null,
+            branch: branch ?? null,
+            version: version ?? null,
+        });
+    }
     async dropIndex({ table, name, namespace, branch }) {
         await this.room.invoke({
             toolkit: "dataset",

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@meshagent/meshagent",
-    "version": "0.39.5",
+    "version": "0.39.6",
     "description": "Meshagent Client",
     "homepage": "https://github.com/meshagent/meshagent-ts",
     "scripts": {