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
[](https://www.youtube.com/live/GHH50rDlLn0?si=-i3hPYq1o6H271_z)
6
6
7
7
## Overview
8
8
@@ -15,26 +15,10 @@ With Foundry Local, you gain:
15
15
-**Cost Efficiency**: No cloud compute costs incurred.
16
16
-**Offline Access**: AI capabilities even without internet connectivity.
17
17
18
-
> [!NOTE]
19
-
>
20
-
> **Hackathon Award Category: Offline-Ready AI Award**
21
-
>
22
-
> This quest mapped to our Offline-Ready AI Award category, which recognized the best performing AI solutions with standout offline capabilities (local inference).
23
-
>
24
-
> Winning solutions featured apps that:
25
-
>
26
-
> - Processed sensitive data entirely on-device.
27
-
> - Used Foundry Local for reasoning and cloud storage services for optional sync or analytics.
28
-
>
29
-
> Winners demonstrated:
30
-
>
31
-
> -**Privacy** (no sensitive data leaving the device).
32
-
> -**Latency** optimization using local inference.
33
-
>
34
-
> Checkout some Foundry Local + JavaScript projects: -
35
-
>
36
-
> 1.[Interview Doctor](https://github.com/leestott/interview-doctor-js) - AI-powered interview preparation assistant that runs **100% offline** on your local machine
37
-
> 1.[Gas Field Support Agent](https://github.com/leestott/local-rag) - A fully offline, on-device Retrieval-Augmented Generation (RAG) support agent for gas field inspection and maintenance engineers.
18
+
Here are two Foundry Local and JavaScript projects you can explore:
19
+
20
+
1.[Interview Doctor](https://github.com/leestott/interview-doctor-js) - An AI-powered interview preparation assistant that runs entirely on your local machine.
21
+
2.[Gas Field Support Agent](https://github.com/leestott/local-rag) - An offline, on-device RAG support agent for gas-field inspection and maintenance engineers.
38
22
39
23
First step, **installation instructions** for your OS:
40
24
@@ -196,21 +180,17 @@ Once configured, GitHub Copilot Chat will use your local Foundry model for all r
196
180
|**Qwen models**| Multilingual code generation |
197
181
|**GPT models**| Advanced capabilities and broad compatibility |
198
182
199
-
> [!NOTE]
200
-
>
201
-
> For the Offline-Ready AI Award, using GitHub Copilot with Foundry Local demonstrates a powerful offline development workflow. Highlight this capability in your submission!
202
-
203
183
### 3. Software Development Kits (SDKs)
204
184
205
185
Foundry Local provides SDKs to programmatically send requests to the local management service. Since the endpoint is dynamically allocated each time the service starts, the SDK handles endpoint discovery and management for you (control plane).
206
186
207
187
#### Step 1: Initialize New Project
208
188
209
-
Create a parent folder for your Build-a-thon projects and navigate into it:
189
+
Create a parent folder for your JavaScript AI projects and navigate into it:
210
190
211
191
```bash
212
-
mkdir buildathon
213
-
cdbuildathon
192
+
mkdir javascript-ai-quests
193
+
cdjavascript-ai-quests
214
194
```
215
195
216
196
Create a new folder for this quest, navigate into it and initialize a Node.js project:
@@ -394,7 +374,7 @@ Example UI:
394
374
395
375
## Stay connected
396
376
397
-
Have a question, project, or insight to share? Join the [Build-a-thon Discord channel](https://aka.ms/JSAIonDiscord)
377
+
Have a question, project, or insight to share? Join the [JavaScript AI community on Discord](https://aka.ms/JSAIonDiscord).
[](https://www.youtube.com/live/8ded7UAdPTA?si=UwncP4b4tIs1dwLS)
6
6
7
7
# Building AI Agents E2E On Microsoft Foundry
8
8
@@ -59,9 +59,7 @@ To complete the quest, you will need:
59
59
>
60
60
> The **Foundry SDK is under active development**.
61
61
>
62
-
> If you encounter any blockers, unclear steps, or have suggestions to improve the developer experience, please open an issue using the [provided issue template](https://github.com/microsoft/microsoft-foundry-e2e-js/issues/new/choose) - *Bug report/ Build-a-thon Feedback*. Make sure to **apply the label** that matches the step you were on when the issue occurred.
63
-
>
64
-
>
62
+
> If you encounter a blocker or unclear instruction in the linked sample, open an issue in the [Microsoft Foundry end-to-end JavaScript repository](https://github.com/microsoft/microsoft-foundry-e2e-js/issues/new/choose). Include the task you were completing and enough environment detail to reproduce the problem.
65
63
66
64
The repoistory is setup with a `devcontainer.json` that installs all the required depdencencies and gives you access to a free tier of GitHub Copilot.
67
65
@@ -90,9 +88,9 @@ This quest has a lot of steps to complete, many of which cannot be completed in
90
88
**Explore the rest at your own pace later**. Create a fork of the repo, then launch a new Codespaces instance on that fork and work through the steps in order. Try customizing the data to suit a different scenario. Or change the code to try a different evaluator or attack strategy for red-teaming. Use your fork as a sandbox for building a deeper understanding of these capabilities with hands-on experiments.
91
89
92
90
93
-
### Return to the Build-a-thon
91
+
### Continue learning
94
92
95
-
Once you have completed this quest and get an intuitive sense for end-to-end development with Microsoft Foundry, return to the main Build-a-thon repository to continue with the next quests.
93
+
When you have an intuitive sense of end-to-end development with Microsoft Foundry, return to the [Quest directory](https://github.com/Azure-Samples/JavaScript-AI-Buildathon#the-learning-path) to choose what to explore next.
0 commit comments