Skip to content

Commit e41e6e6

Browse files
siddharthsambharia-portkeysiddharthsambharia-portkey
authored
Merge pull request #300 from Portkey-AI/azure-guardrails
azure guardrails on portkey
2 parents 8560880 + 03b2eb0 commit e41e6e6

File tree

3 files changed

+227
-1
lines changed

3 files changed

+227
-1
lines changed

‎docs.json

Lines changed: 6 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -101,9 +101,10 @@
101101
"pages": [
102102
"product/guardrails",
103103
"product/guardrails/list-of-guardrail-checks",
104+
"product/guardrails/azure-guardrails",
104105
"product/guardrails/pii-redaction",
105106
"product/guardrails/patronus-ai",
106-
"product/guardrails/bedrock-guardrials",
107+
"product/guardrails/bedrock-guardrails",
107108
"product/guardrails/lasso",
108109
"product/guardrails/aporia",
109110
"product/guardrails/pillar",
@@ -2019,6 +2020,10 @@
20192020
"source": "/welcome/integration-guides/gemini",
20202021
"destination": "/integrations/llms/gemini"
20212022
},
2023+
{
2024+
"source": "product/guardrails/bedrock-guardrials",
2025+
"destination": "product/guardrails/bedrock-guardrails"
2026+
},
20222027
{
20232028
"source": "/api-reference/prompts",
20242029
"destination": "/api-reference/inference-api/prompts"

‎product/guardrails/azure-guardrails.mdx

Lines changed: 221 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,221 @@
1+
---
2+
title: "Azure Guardrails"
3+
description: "Integrate Microsoft Azure's powerful content moderation services & PII guardrails with Portkey"
4+
---
5+
6+
Microsoft Azure offers robust content moderation and PII readaction services that can now be seamlessly integrated with Portkey's guardrails ecosystem. This integration supports two powerful Azure services:
7+
8+
<CardGroup cols={2}>
9+
<Card title="Azure Content Safety" icon="shield-check">
10+
A comprehensive content moderation service that detects harmful content including hate speech, violence, sexual content, and self-harm references in text.
11+
</Card>
12+
<Card title="Azure PII Detection" icon="user-lock">
13+
Advanced detection of personally identifiable information (PII) and protected health information (PHI) to safeguard sensitive data.
14+
</Card>
15+
</CardGroup>
16+
17+
## Setting Up Azure Guardrails
18+
19+
Follow these steps to integrate Azure's content moderation services with Portkey:
20+
21+
### 1. Configure Azure Authentication
22+
23+
Navigate to the **Integrations** page under **Settings** to set up your Azure credentials. You can authenticate using three different methods:
24+
25+
- **API Key** - Uses a simple API key for authentication
26+
- **Entra** (formerly Azure AD) - Uses Azure Active Directory authentication
27+
- **Managed** - Uses managed identity authentication within Azure
28+
29+
Each authentication method requires specific credentials from your Azure account:
30+
31+
<Tabs>
32+
<Tab title="API Key">
33+
- **Resource Name**: Your Azure resource name
34+
- **API Key**: Your Azure API key
35+
</Tab>
36+
<Tab title="Entra">
37+
- **Resource Name**: Your Azure resource name
38+
- **Client ID**: Your Azure client ID
39+
- **Client Secret**: Your client secret
40+
- **Tenant ID**: Your Azure tenant ID
41+
</Tab>
42+
<Tab title="Managed">
43+
- **Resource Name**: Your Azure resource name
44+
- **Client ID**: Your Azure client ID (for managed identity)
45+
</Tab>
46+
</Tabs>
47+
48+
### 2. Create Azure Guardrail Checks
49+
50+
Once authentication is set up, you can add Azure guardrail checks to your Portkey workflow:
51+
52+
1. Navigate to the **Guardrails** page
53+
2. Search for either **Azure Content Safety** or **Azure PII Detection**
54+
3. Click **Add** to configure your chosen guardrail
55+
4. Configure the specific settings for your guardrail
56+
5. Save your configuration and create the guardrail
57+
58+
<Note>
59+
Guardrail Actions allow you to orchestrate your guardrails logic. You can learn more about them [here](/product/guardrails#there-are-6-types-of-guardrail-actions)
60+
</Note>
61+
62+
## Azure Content Safety
63+
64+
Azure Content Safety analyzes text for harmful content across several categories.
65+
66+
**Configuration Options**
67+
68+
| Parameter | Description | Values |
69+
|-----------|-------------|--------|
70+
| Blocklist Names | Custom Blocklist names from your azure setup | `blocklist-1`, `blocklist-2`, `blocklist-3` |
71+
| API Version | Azure Content Safety API version | Default: `2024-09-01` |
72+
| Severity | Minimum severity threshold for flagging content | `2`, `4`, `6`, or `8` |
73+
| Categories | Content categories to monitor | Hate, SelfHarm, Sexual, Violence |
74+
| Timeout | Maximum time in milliseconds for the check | Default: `5000` |
75+
76+
### Using Blocklists
77+
78+
Blocklists allow you to define custom terms or patterns to be flagged. You'll need to create Content Safety blocklists in your Azure account first, then reference them in the Blocklist Names field.
79+
80+
<Info>
81+
For more information on Azure Content Safety blocklists, visit the [official documentation](https://learn.microsoft.com/en-us/azure/ai-services/content-safety/quickstart-blocklist).
82+
</Info>
83+
84+
## Azure PII Detection
85+
86+
Azure PII Detection identifies and can help protect personal and health-related information in your content.
87+
88+
**Configuration Options**
89+
90+
| Parameter | Description | Values |
91+
|-----------|-------------|--------|
92+
| Domain | The type of sensitive information to detect | `none` (both PII and PHI) or `phi` (only PHI) |
93+
| API Version | Azure PII Detection API version | Default: `2024-11-01` |
94+
| Model Version | Version of the detection model to use | Default: `latest` |
95+
| Redact | Option to redact detected information | `true` or `false` |
96+
| Timeout | Maximum time in milliseconds for the check | Default: `5000` |
97+
98+
## Add Guardrail ID to a Config and Make Your Request
99+
100+
* When you save a Guardrail, you'll get an associated Guardrail ID - add this ID to the `input_guardrails` or `output_guardrails` params in your Portkey Config
101+
* Create these Configs in Portkey UI, save them, and get an associated Config ID to attach to your requests. [More here](/product/ai-gateway/configs).
102+
103+
Here's an example configuration:
104+
105+
```json
106+
{
107+
"input_guardrails": ["guardrails-id-xxx", "guardrails-id-yyy"],
108+
"output_guardrails": ["guardrails-id-xxx", "guardrails-id-yyy"]
109+
}
110+
```
111+
112+
<Tabs>
113+
<Tab title="NodeJS">
114+
```js
115+
const portkey = new Portkey({
116+
apiKey: "PORTKEY_API_KEY",
117+
config: "pc-***" // Supports a string config id or a config object
118+
});
119+
```
120+
</Tab>
121+
<Tab title="Python">
122+
```py
123+
portkey = Portkey(
124+
api_key="PORTKEY_API_KEY",
125+
config="pc-***" # Supports a string config id or a config object
126+
)
127+
```
128+
</Tab>
129+
<Tab title="OpenAI NodeJS">
130+
```js
131+
const openai = new OpenAI({
132+
apiKey: 'OPENAI_API_KEY',
133+
baseURL: PORTKEY_GATEWAY_URL,
134+
defaultHeaders: createHeaders({
135+
apiKey: "PORTKEY_API_KEY",
136+
config: "CONFIG_ID"
137+
})
138+
});
139+
```
140+
</Tab>
141+
<Tab title="OpenAI Python">
142+
```py
143+
client = OpenAI(
144+
api_key="OPENAI_API_KEY", # defaults to os.environ.get("OPENAI_API_KEY")
145+
base_url=PORTKEY_GATEWAY_URL,
146+
default_headers=createHeaders(
147+
provider="openai",
148+
api_key="PORTKEY_API_KEY", # defaults to os.environ.get("PORTKEY_API_KEY")
149+
config="CONFIG_ID"
150+
)
151+
)
152+
```
153+
</Tab>
154+
<Tab title="cURL">
155+
```sh
156+
curl https://api.portkey.ai/v1/chat/completions \
157+
-H "Content-Type: application/json" \
158+
-H "Authorization: Bearer $OPENAI_API_KEY" \
159+
-H "x-portkey-api-key: $PORTKEY_API_KEY" \
160+
-H "x-portkey-config: $CONFIG_ID" \
161+
-d '{
162+
"model": "gpt-3.5-turbo",
163+
"messages": [{
164+
"role": "user",
165+
"content": "Hello!"
166+
}]
167+
}'
168+
```
169+
</Tab>
170+
</Tabs>
171+
172+
For more, refer to the [Config documentation](/product/ai-gateway/configs).
173+
174+
175+
176+
177+
## Monitoring and Logs
178+
179+
All guardrail actions and verdicts are visible in your Portkey logs, allowing you to:
180+
181+
- Track which content has been flagged
182+
- See guardrail verdicts and actions
183+
- Monitor the performance of your content moderation pipeline
184+
185+
186+
<Frame>
187+
<img src="/images/product/ai-gateway/ai-34.gif"/>
188+
</Frame>
189+
190+
191+
## Using Azure Guardrails - Scenarios
192+
193+
After setting up your guardrails, there are different ways to use them depending on your security requirements:
194+
195+
### Detect and Monitor Only
196+
197+
To simply detect but not block content:
198+
- Configure your guardrail actions without enabling "Deny"
199+
- Monitor the guardrail results in your Portkey logs
200+
- If any issues are detected, the response will include a `hook_results` object with details
201+
202+
### Redact PII Automatically
203+
204+
To automatically remove sensitive information:
205+
- Enable the `Redact` option for Azure PII Detection
206+
- When PII is detected, it will be automatically redacted and replaced with standardized identifiers
207+
- The response will include a `transformed` flag set to `true` in the results
208+
209+
### Block Harmful Content
210+
211+
To completely block requests that violate your policies:
212+
- Enable the `Deny` option in the guardrails action tab
213+
- If harmful content is detected, the request will fail with an appropriate status code
214+
- You can customize denial messages to provide guidance to users
215+
216+
217+
---
218+
219+
## Need Support?
220+
221+
If you encounter any issues with Azure Guardrails, please reach out to our support team through the [Portkey community forum](https://discord.gg/portkey-llms-in-prod-1143393887742861333).

‎product/guardrails/bedrock-guardrials.mdx renamed to ‎product/guardrails/bedrock-guardrails.mdx

File renamed without changes.

0 commit comments

Comments
 (0)