You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/LLMs.html
+15-15Lines changed: 15 additions & 15 deletions
Original file line number
Diff line number
Diff line change
@@ -258,21 +258,21 @@
258
258
<h1>Large Language Models (LLMs) features<aclass="headerlink" href="#large-language-models-llms-features" title="Link to this heading">¶</a></h1>
259
259
<p>Our library offers two ways to interact with Large Language Models (LLMs):</p>
260
260
<olclass="arabic simple">
261
-
<li><p><strong><codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> (Recommended)</strong>: A lightweight server that provides a simple, high-level interface for common task-related operations like cloning, transforming, and localizing PsyFlow tasks. This is the easiest and recommended way to get started.</p></li>
262
-
<li><p><strong>Built-in <codeclass="docutils literal notranslate"><spanclass="pre">LLMClient</span></code> (Lower-Level)</strong>: A minimal wrapper around LLM provider SDKs (Gemini, Deepseek) for more direct control. This is suitable for developers who need to customize the LLM interaction beyond the scope of <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code>.</p></li>
261
+
<li><p><strong><codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> (Recommended)</strong>: A lightweight server that provides a simple, high-level interface for common task-related operations like cloning, transforming, and localizing PsyFlow tasks. This is the easiest and recommended way to get started.</p></li>
262
+
<li><p><strong>Built-in <codeclass="docutils literal notranslate"><spanclass="pre">LLMClient</span></code> (Lower-Level)</strong>: A minimal wrapper around LLM provider SDKs (Gemini, Deepseek) for more direct control. This is suitable for developers who need to customize the LLM interaction beyond the scope of <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code>.</p></li>
263
263
</ol>
264
264
<p><strong>Why It Matters</strong>: Large Language Models (LLMs) significantly enhance the usability and reproducibility of cognitive task development. They enable researchers to translate configuration files for localization, generate detailed documentation from code, and prototype or refine task variants using natural language—all while avoiding repetitive formatting work. By integrating LLMs directly into the PsyFlow ecosystem, we accelerate development, promote clearer communication, and expand accessibility for both developers and collaborators.</p>
265
265
<hrclass="docutils" />
266
-
<sectionid="psyflow-mcp-recommended">
267
-
<h2>1. <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> (Recommended)<aclass="headerlink" href="#psyflow-mcp-recommended" title="Link to this heading">¶</a></h2>
268
-
<p><codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> is a lightweight server that simplifies the use of LLMs for managing PsyFlow tasks. It exposes a set of tools that can be easily integrated with LLM agents like the Gemini CLI or Cursor.</p>
266
+
<sectionid="taskbeacon-mcp-recommended">
267
+
<h2>1. <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> (Recommended)<aclass="headerlink" href="#taskbeacon-mcp-recommended" title="Link to this heading">¶</a></h2>
268
+
<p><codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> is a lightweight server that simplifies the use of LLMs for managing PsyFlow tasks. It exposes a set of tools that can be easily integrated with LLM agents like the Gemini CLI or Cursor.</p>
269
269
<p><strong>How It Works</strong></p>
270
-
<p>The <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> server acts as a bridge between the user’s natural language prompts and the underlying PsyFlow task management functions. The workflow is as follows:</p>
270
+
<p>The <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> server acts as a bridge between the user’s natural language prompts and the underlying PsyFlow task management functions. The workflow is as follows:</p>
271
271
<p><imgalt="Illustration of the MCP Workflow" src="_images/LLM_flowchart.png" /></p>
272
272
<olclass="arabic simple">
273
273
<li><p><strong>User Prompt</strong>: The user provides a prompt describing the desired action (e.g., “Create an SST task with sound-based stop signals” or “Give me a French version of the SST task”).</p></li>
274
-
<li><p><strong>LLM</strong>: The LLM interprets the prompt and selects the appropriate tool from <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code>.</p></li>
275
-
<li><p><strong>MCP (Model Context Protocol)</strong>: The <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> server executes the requested tool, which may involve:</p>
274
+
<li><p><strong>LLM</strong>: The LLM interprets the prompt and selects the appropriate tool from <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code>.</p></li>
275
+
<li><p><strong>MCP (Model Context Protocol)</strong>: The <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> server executes the requested tool, which may involve:</p>
276
276
<ulclass="simple">
277
277
<li><p><codeclass="docutils literal notranslate"><spanclass="pre">build_task</span></code>: Cloning a task template and preparing it for modification.</p></li>
278
278
<li><p><codeclass="docutils literal notranslate"><spanclass="pre">localize</span></code>: Translating a task’s configuration file.</p></li>
<p>The easiest way to use <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> is with <codeclass="docutils literal notranslate"><spanclass="pre">uvx</span></code>, which handles the installation and execution in a single command.</p>
286
+
<p>The easiest way to use <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> is with <codeclass="docutils literal notranslate"><spanclass="pre">uvx</span></code>, which handles the installation and execution in a single command.</p>
<spanclass="w"></span><spanclass="nt">"description"</span><spanclass="p">:</span><spanclass="w"></span><spanclass="s2">"Local FastMCP server for PsyFlow task operations. Uses uvx for automatic setup."</span><spanclass="p">,</span>
<p>With this setup, your LLM agent can now use the <codeclass="docutils literal notranslate"><spanclass="pre">psyflow_mcp</span></code> tools. For more details, refer to the <aclass="reference external" href="https://github.com/TaskBeacon/psyflow-mcp"><codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code> documentation</a>.</p>
307
+
<p>With this setup, your LLM agent can now use the <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> tools. For more details, refer to the <aclass="reference external" href="https://github.com/TaskBeacon/taskbeacon-mcp"><codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code> documentation</a>.</p>
<h2>3. LLMs-Powered Task Documentation<aclass="headerlink" href="#llms-powered-task-documentation" title="Link to this heading">¶</a></h2>
402
402
<p>Our platform leverages Large Language Models (LLMs) to automatically generate human-readable documentation for cognitive tasks. This feature is designed to help developers, collaborators, and reviewers quickly understand the structure and parameters of a task—without having to dig through source code.</p>
403
-
<p>While this can be done manually with the <codeclass="docutils literal notranslate"><spanclass="pre">LLMClient</span></code>, it is more easily accomplished using the <codeclass="docutils literal notranslate"><spanclass="pre">build_task</span></code> tool in <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code>.</p>
403
+
<p>While this can be done manually with the <codeclass="docutils literal notranslate"><spanclass="pre">LLMClient</span></code>, it is more easily accomplished using the <codeclass="docutils literal notranslate"><spanclass="pre">build_task</span></code> tool in <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code>.</p>
404
404
<p>Our <codeclass="docutils literal notranslate"><spanclass="pre">LLMClient</span></code> includes a powerful <codeclass="docutils literal notranslate"><spanclass="pre">task2doc()</span></code> utility that lets you <strong>automatically generate a detailed <codeclass="docutils literal notranslate"><spanclass="pre">README.md</span></code></strong> file for any PsyFlow-based cognitive task.</p>
405
405
<p><codeclass="docutils literal notranslate"><spanclass="pre">task2doc()</span></code> analyzes four types of files:</p>
<h2>4. LLMs-Powered Localization<aclass="headerlink" href="#llms-powered-localization" title="Link to this heading">¶</a></h2>
441
441
<p>The <codeclass="docutils literal notranslate"><spanclass="pre">LLMClient</span></code> also supports automatic translation of task configurations using the <codeclass="docutils literal notranslate"><spanclass="pre">translate_config()</span></code> method. This localization feature enables your task templates to be easily adapted into other languages while preserving placeholder tokens and formatting. By combining this with PsyFlow’s localization-ready structure, you can easily localize tasks for global deployment.</p>
442
-
<p>This is more easily accomplished using the <codeclass="docutils literal notranslate"><spanclass="pre">localize</span></code> tool in <codeclass="docutils literal notranslate"><spanclass="pre">psyflow-mcp</span></code>.</p>
442
+
<p>This is more easily accomplished using the <codeclass="docutils literal notranslate"><spanclass="pre">localize</span></code> tool in <codeclass="docutils literal notranslate"><spanclass="pre">taskbeacon-mcp</span></code>.</p>
443
443
<p><codeclass="docutils literal notranslate"><spanclass="pre">translate_config()</span></code> translate the following content in configuration:</p>
Copy file name to clipboardExpand all lines: docs/_sources/LLMs.md.txt
+13-13Lines changed: 13 additions & 13 deletions
Original file line number
Diff line number
Diff line change
@@ -2,28 +2,28 @@
2
2
3
3
Our library offers two ways to interact with Large Language Models (LLMs):
4
4
5
-
1. **`psyflow-mcp` (Recommended)**: A lightweight server that provides a simple, high-level interface for common task-related operations like cloning, transforming, and localizing PsyFlow tasks. This is the easiest and recommended way to get started.
6
-
2. **Built-in `LLMClient` (Lower-Level)**: A minimal wrapper around LLM provider SDKs (Gemini, Deepseek) for more direct control. This is suitable for developers who need to customize the LLM interaction beyond the scope of `psyflow-mcp`.
5
+
1. **`taskbeacon-mcp` (Recommended)**: A lightweight server that provides a simple, high-level interface for common task-related operations like cloning, transforming, and localizing PsyFlow tasks. This is the easiest and recommended way to get started.
6
+
2. **Built-in `LLMClient` (Lower-Level)**: A minimal wrapper around LLM provider SDKs (Gemini, Deepseek) for more direct control. This is suitable for developers who need to customize the LLM interaction beyond the scope of `taskbeacon-mcp`.
7
7
8
8
**Why It Matters**: Large Language Models (LLMs) significantly enhance the usability and reproducibility of cognitive task development. They enable researchers to translate configuration files for localization, generate detailed documentation from code, and prototype or refine task variants using natural language—all while avoiding repetitive formatting work. By integrating LLMs directly into the PsyFlow ecosystem, we accelerate development, promote clearer communication, and expand accessibility for both developers and collaborators.
9
9
10
10
---
11
11
12
-
### 1. `psyflow-mcp` (Recommended)
12
+
### 1. `taskbeacon-mcp` (Recommended)
13
13
14
-
`psyflow-mcp` is a lightweight server that simplifies the use of LLMs for managing PsyFlow tasks. It exposes a set of tools that can be easily integrated with LLM agents like the Gemini CLI or Cursor.
14
+
`taskbeacon-mcp` is a lightweight server that simplifies the use of LLMs for managing PsyFlow tasks. It exposes a set of tools that can be easily integrated with LLM agents like the Gemini CLI or Cursor.
15
15
16
16
**How It Works**
17
17
18
-
The `psyflow-mcp` server acts as a bridge between the user's natural language prompts and the underlying PsyFlow task management functions. The workflow is as follows:
18
+
The `taskbeacon-mcp` server acts as a bridge between the user's natural language prompts and the underlying PsyFlow task management functions. The workflow is as follows:
19
19
20
20
21
21
22
22
23
23
24
24
1. **User Prompt**: The user provides a prompt describing the desired action (e.g., "Create an SST task with sound-based stop signals" or "Give me a French version of the SST task").
25
-
2. **LLM**: The LLM interprets the prompt and selects the appropriate tool from `psyflow-mcp`.
26
-
3. **MCP (Model Context Protocol)**: The `psyflow-mcp` server executes the requested tool, which may involve:
25
+
2. **LLM**: The LLM interprets the prompt and selects the appropriate tool from `taskbeacon-mcp`.
26
+
3. **MCP (Model Context Protocol)**: The `taskbeacon-mcp` server executes the requested tool, which may involve:
27
27
* `build_task`: Cloning a task template and preparing it for modification.
28
28
* `localize`: Translating a task's configuration file.
29
29
* `download_task`: Fetching a task from the registry.
@@ -32,7 +32,7 @@ The `psyflow-mcp` server acts as a bridge between the user's natural language pr
32
32
33
33
**Quick Start**
34
34
35
-
The easiest way to use `psyflow-mcp` is with `uvx`, which handles the installation and execution in a single command.
35
+
The easiest way to use `taskbeacon-mcp` is with `uvx`, which handles the installation and execution in a single command.
36
36
37
37
1. **Install `uvx`**:
38
38
```bash
@@ -42,17 +42,17 @@ The easiest way to use `psyflow-mcp` is with `uvx`, which handles the installati
42
42
Create a JSON configuration file for your LLM tool (e.g., Gemini CLI) with the following content:
43
43
```json
44
44
{
45
-
"name": "psyflow_mcp",
45
+
"name": "taskbeacon-mcp",
46
46
"type": "stdio",
47
47
"description": "Local FastMCP server for PsyFlow task operations. Uses uvx for automatic setup.",
48
48
"isActive": true,
49
49
"command": "uvx",
50
50
"args": [
51
-
"psyflow_mcp"
51
+
"taskbeacon-mcp"
52
52
]
53
53
}
54
54
```
55
-
With this setup, your LLM agent can now use the `psyflow_mcp` tools. For more details, refer to the [`psyflow-mcp` documentation](https://github.com/TaskBeacon/psyflow-mcp).
55
+
With this setup, your LLM agent can now use the `taskbeacon-mcp` tools. For more details, refer to the [`taskbeacon-mcp` documentation](https://github.com/TaskBeacon/taskbeacon-mcp).
Our platform leverages Large Language Models (LLMs) to automatically generate human-readable documentation for cognitive tasks. This feature is designed to help developers, collaborators, and reviewers quickly understand the structure and parameters of a task—without having to dig through source code.
152
152
153
-
While this can be done manually with the `LLMClient`, it is more easily accomplished using the `build_task` tool in `psyflow-mcp`.
153
+
While this can be done manually with the `LLMClient`, it is more easily accomplished using the `build_task` tool in `taskbeacon-mcp`.
154
154
155
155
Our `LLMClient` includes a powerful `task2doc()` utility that lets you **automatically generate a detailed `README.md`** file for any PsyFlow-based cognitive task.
156
156
@@ -192,7 +192,7 @@ This automatic documentation feature reduces the burden on developers, promotes
192
192
193
193
The `LLMClient` also supports automatic translation of task configurations using the `translate_config()` method. This localization feature enables your task templates to be easily adapted into other languages while preserving placeholder tokens and formatting. By combining this with PsyFlow’s localization-ready structure, you can easily localize tasks for global deployment.
194
194
195
-
This is more easily accomplished using the `localize` tool in `psyflow-mcp`.
195
+
This is more easily accomplished using the `localize` tool in `taskbeacon-mcp`.
196
196
197
197
`translate_config()` translate the following content in configuration:
0 commit comments