Unify

https://docs.unify.ai/
Admin
Unify is a platform that provides unified access to multiple AI models and services through a single...
Tokens:157,684
Snippets:1,157
Trust Score:9
Update:10 months ago
Show doc for...
Tokens:
Raw
### Python: Unify AI Quick Start Evaluation

Source: https://docs.unify.ai/basics/quickstart

This Python script demonstrates how to initialize a Unify AI project, set up a language model agent, define test cases, and run an evaluation. It logs questions, agent responses, and performance scores for each test case within an experiment.

```python
import unify
from random import randint, choice

# initialize project
unify.activate("Maths Assistant")

# build agent
client = unify.Unify("o3-mini@openai", traced=True)
client.set_system_message(
    "You are a helpful maths assistant, "
    "tasked with adding and subtracting integers."
)

# add test cases
qs = [
    f"{randint(0, 100)} {choice(['+', '-'])} {randint(0, 100)}"
    for i in range(10)
]

# define evaluator
@unify.traced
def evaluate_response(question: str, response: str) -> float:
    correct_answer = eval(question)
    try:
        response_int = int(
            "".join(
                [
                    c for c in response.split(" ")[-1]
                    if c.isdigit()
                ]
            ),
        )
        return float(correct_answer == response_int)
    except ValueError:
        return 0.

# define evaluation
@unify.traced
def evaluate(q: str):
    response = client.copy().generate(q)
    score = evaluate_response(q, response)
    unify.log(
        question=q,
        response=response,
        score=score
    )

# execute + log your evaluation
with unify.Experiment():
    unify.map(evaluate, qs)
```

--------------------------------

### Unify AI Vision Example (Fails)

Source: https://docs.unify.ai/universal_api/queries

An example attempting to use vision features with the Unify AI client, showing that it fails with Anthropic's native formatting. This highlights the need to adopt OpenAI's format for vision.

```python
import unify
import base64
import httpx

image_url = "https://raw.githubusercontent.com/"
"unifyai/unifyai.github.io/refs/heads/main/"
"img/externally_linked/ant.jpg"
image_media_type = "image/jpeg"
image_data = base64.b64encode(httpx.get(image_url).content).decode("utf-8")

client = unify.Unify("claude-3-sonnet@anthropic")
response = client.generate(
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What’s in this image?"},
                {
                    "type": "image",
                    "source": {
                        "type": "base64",
                        "media_type": image_media_type,
                        "data": image_data,
                    },
                },
            ],
        }
    ],
    max_completion_tokens=300,
)
print(response)
```

--------------------------------

### Tool Use Example

Source: https://docs.unify.ai/universal_api/queries

Demonstrates how to use the tool use feature with the Unify AI API, adhering to the OpenAI standard. This example shows how to call an Anthropic model using OpenAI's function calling interface.

```APIDOC
## POST /api/chat/completions (Tool Use)

### Description
This endpoint allows for tool use with various LLM providers, unified under the OpenAI standard. It demonstrates how to integrate function calling capabilities into your LLM interactions.

### Method
POST

### Endpoint
/api/chat/completions

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **model** (string) - Required - The model to use for generation.
- **messages** (array) - Required - A list of messages comprising the conversation.
- **tools** (array) - Optional - A list of tools the model may call.
- **tool_choice** (string or object) - Optional - Controls how the tool is chosen.
- **message_content_only** (boolean) - Optional - If true, returns only the content of the message.

### Request Example
```json
{
  "model": "claude-3.5-sonnet@aws-bedrock",
  "messages": [
    {
      "role": "user",
      "content": "What is the current temperature of New York, San Francisco and Chicago?"
    }
  ],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_current_weather",
        "description": "Get the weather in a given location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {
              "type": "string",
              "description": "City and state, ie New York, NY"
            },
            "unit": {
              "type": "string",
              "enum": ["celsius", "fahrenheit"]
            }
          }
        }
      }
    }
  ],
  "tool_choice": "auto",
  "message_content_only": false
}
```

### Response
#### Success Response (200)
- **choices** (array) - The list of generated choices.
  - **message** (object) - The message content.
    - **role** (string) - The role of the author.
    - **tool_calls** (array) - The tool calls to be made.
      - **id** (string) - The ID of the tool call.
      - **type** (string) - The type of the tool call.
      - **function** (object) - The function to call.
        - **name** (string) - The name of the function.
        - **arguments** (string) - The arguments to pass to the function.

#### Response Example
```json
{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "tool_calls": [
          {
            "id": "call_abc123",
            "type": "function",
            "function": {
              "name": "get_current_weather",
              "arguments": "{\"location\": \"New York, NY\", \"unit\": \"celsius\"}"
            }
          },
          {
            "id": "call_def456",
            "type": "function",
            "function": {
              "name": "get_current_weather",
              "arguments": "{\"location\": \"San Francisco, CA\", \"unit\": \"celsius\"}"
            }
          },
          {
            "id": "call_ghi789",
            "type": "function",
            "function": {
              "name": "get_current_weather",
              "arguments": "{\"location\": \"Chicago, IL\", \"unit\": \"celsius\"}"
            }
          }
        ]
      }
    }
  ]
}
```
```

--------------------------------

### Chat Completion Request Example (cURL)

Source: https://docs.unify.ai/api-reference/llm_queries/chat_completions

This example demonstrates how to make a chat completion request using cURL. It includes common parameters like messages, model, max_tokens, stop sequences, temperature, and tool_choice.

```bash
curl --request POST \
  --url 'https://api.unify.ai/v0/chat/completions' \
  --header "Authorization: Bearer $UNIFY_KEY" \
  --header 'Content-Type: application/json' \
  --data '{
      "messages": [
          {
              "content": "Tell me a joke",
              "role": "user"
          }
      ],
      "model": "gpt-4o-mini@openai",
      "max_tokens": 1024,
      "stop": [
          "The End.",
          " is the answer."
      ],
      "stream": false,
      "temperature": 0.9,
      "frequency_penalty": 1.5,
      "logit_bias": {
          "0": 10,
          "1": -75,
          "2": 90
      },
      "logprobs": false,
      "top_logprobs": 15,
      "n": 15,
      "presence_penalty": -1.1,
      "response_format": "{ \"type\": \"json_mode\"}",
      "seed": 11,
      "stream_options": [
          true,
          "include_usage"
      ],
      "top_p": 0.5,
      "tool_choice": "{\"type\": \"function\", \"function\": {\"name\": \"my_function\"}}",
      "parallel_tool_calls": true,
      "user": "some_user",
      "signature": "python",
      "use_custom_keys": true,
      "tags": [
          "user123",
          "CompanyABC"
      ],
      "drop_params": true
  }'
```

--------------------------------

### Anthropic Vision Example

Source: https://docs.unify.ai/universal_api/queries

An example demonstrating how to use vision features with the Anthropic client, including sending base64 encoded image data. This example uses Anthropic's native formatting.

```python
import anthropic
import base64
import httpx

image_url = "https://raw.githubusercontent.com/"
"unifyai/unifyai.github.io/refs/heads/main/"
"img/externally_linked/ant.jpg"
image_media_type = "image/jpeg"
image_data = base64.b64encode(httpx.get(image_url).content).decode("utf-8")

client = anthropic.Anthropic()
response = client.messages.create(
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What’s in this image?"},
                {
                    "type": "image",
                    "source": {
                        "type": "base64",
                        "media_type": image_media_type,
                        "data": image_data,
                    },
                },
            ],
        }
    ],
    max_completion_tokens=300,
    model="claude-3-opus-20240229"
)
print(response)
```

--------------------------------

### Fetch UnifyAI Documentation File

Source: https://docs.unify.ai/api-reference/assistant_management/create_a_new_assistant

This instruction guides users on how to access additional documentation files for the UnifyAI project. It specifies the file name and the URL where it can be retrieved.

```text
To find navigation and other pages in this documentation, fetch the llms.txt file at: https://docs.unify.ai/llms.txt
```

--------------------------------

### Get Default Tools

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/MultiUnify

Retrieves the default tools configured for the Unify AI client.

```APIDOC
## GET /unify/tools

### Description
Get the default tools, if set.

### Method
GET

### Endpoint
/unify/tools

### Parameters
None

### Request Example
None

### Response
#### Success Response (200)
- **tools** (Iterable[ChatCompletionToolParam]) - The list of default tools.

#### Response Example
{
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "example_function"
      }
    }
  ]
}
```

--------------------------------

### GET /credits

Source: https://docs.unify.ai/python/functions/universal_api/utils/credits/get_credits

Retrieves the remaining credits in the user account, in USD.

```APIDOC
## GET /credits

### Description
Returns the credits remaining in the user account, in USD.

### Method
GET

### Endpoint
/credits

### Parameters
#### Query Parameters
- **api_key** (string) - Optional - If specified, unify API key to be used.

### Request Example
```json
{
  "api_key": "YOUR_API_KEY"
}
```

### Response
#### Success Response (200)
- **credits** (float) - The credits remaining in USD.

#### Response Example
```json
{
  "credits": 10.50
}
```

#### Error Response
- **ValueError**: If there was an HTTP error.
```

--------------------------------

### POST /reset_all

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/AsyncMultiUnify

Resets all base client properties to their default values. This is useful for starting fresh or troubleshooting configuration issues.

```APIDOC
## POST /reset_all

### Description
Reset base client properties to their default values.

### Method
POST

### Endpoint
/reset_all

### Parameters
None

### Request Example
None

### Response
#### Success Response (200)
- **Self** (Self) - This client, useful for chaining inplace calls.

#### Response Example
```json
{
  "message": "Client reset to default values."
}
```
```

--------------------------------

### Get Default Top Logprobs

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/MultiUnify

Retrieves the default top logprobs setting for the Unify AI client.

```APIDOC
## GET /unify/top_logprobs

### Description
Get the default top logprobs, if set.

### Method
GET

### Endpoint
/unify/top_logprobs

### Parameters
None

### Request Example
None

### Response
#### Success Response (200)
- **top_logprobs** (int) - The default top logprobs value.

#### Response Example
{
  "top_logprobs": 5
}
```

--------------------------------

### Get Default Tool Choice

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/MultiUnify

Retrieves the default tool choice configuration for the Unify AI client.

```APIDOC
## GET /unify/tool_choice

### Description
Get the default tool choice, if set.

### Method
GET

### Endpoint
/unify/tool_choice

### Parameters
None

### Request Example
None

### Response
#### Success Response (200)
- **tool_choice** (ChatCompletionToolChoiceOptionParam) - The default tool choice configuration.

#### Response Example
{
  "tool_choice": {
    "type": "auto"
  }
}
```

--------------------------------

### Run ChatBot

Source: https://docs.unify.ai/python/classes/universal_api/chatbot/ChatBot

Starts the chat interaction loop, allowing for conversational interaction with the LLM.

```APIDOC
## run ChatBot

### Description
Starts the chat interaction loop.

### Method
POST

### Endpoint
/websites/unify_ai/chatbot/run

### Parameters
#### Query Parameters
- **show_credits** (boolean) - Optional - Whether to show credit consumption. Defaults to False.
- **show_endpoint** (boolean) - Optional - Whether to show the endpoint used. Defaults to False.

### Response
#### Success Response (200)
- **message** (string) - Chat interaction started.

#### Response Example
{
  "message": "Chat interaction started. Type 'quit' to exit."
}
```

--------------------------------

### Get Default Stream Options

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/AsyncMultiUnify

Retrieves the default stream options, which configure how the stream response is handled.

```APIDOC
## GET /api/config/stream_options

### Description
Get the default stream options, if set.

### Method
GET

### Endpoint
/api/config/stream_options

### Parameters
None

### Request Example
None

### Response
#### Success Response (200)
- **stream_options** (Optional[ChatCompletionStreamOptionsParam]) - The default stream options.

#### Response Example
{
  "stream_options": {"include_usage": false}
}
```

--------------------------------

### POST /v0/assistant/photo/animate

Source: https://docs.unify.ai/api-reference/media/animate_photo

Starts a job to generate an animated video of the assistant using an input image and audio.

```APIDOC
## POST /v0/assistant/photo/animate

### Description
Starts a job to generate an animated video of the assistant using an input image and audio. This action costs credits.

### Method
POST

### Endpoint
https://api.unify.ai/v0/assistant/photo/animate

### Parameters
#### Path Parameters
*None*

#### Query Parameters
*None*

#### Request Body
- **image_url** (string | null) - URL of the input image.
- **image_file** (string | null) - Base64 encoded input image file.
- **audio_url** (string | null) - URL of the input audio.
- **audio_file** (string | null) - Base64 encoded input audio file.
- **seed** (integer | null) - Seed for generating the animation.
- **duration** (integer | null) - Desired duration of the output video in seconds.

### Request Example
```json
{
  "image_url": "<url_to_image>",
  "audio_url": "<url_to_audio>"
}
```

### Response
#### Success Response (202)
- **info** (object) - Information about the animation job.

#### Response Example
```json
{
    "info": {}
}
```

#### Error Response (422)
- **detail** (array) - Array of error details.
  - **loc** (string) - Location of the error.
  - **msg** (string) - Error message.
  - **type** (string) - Error type.

#### Error Response Example
```json
{
  "detail": [
    {
      "loc": ["string"],
      "msg": "string",
      "type": "string"
    }
  ]
}
```
```

--------------------------------

### Get Credit Balance with Unify AI Python Client

Source: https://docs.unify.ai/basics/billing

This snippet shows how to fetch your credit balance using the Unify AI Python client library. It assumes the library is installed and configured with your API key.

```python
import unify
credits = unify.get_credits()
```

--------------------------------

### Retrieve Queries with Filters (cURL)

Source: https://docs.unify.ai/api-reference/usage/get_queries

This snippet demonstrates how to retrieve query history using cURL. It includes authorization via a bearer token and filters results by tags, endpoints, start and end times, page number, and failure status. The endpoint is `https://api.unify.ai/v0/queries`.

```bash
curl --request GET \
  --url 'https://api.unify.ai/v0/queries?tags=my_tag&endpoints=gpt-4o@openai&start_time=2024-07-12%2004:20:32&end_time=2024-08-12%2004:20:32&page_number=1&failures=False' \
  --header "Authorization: Bearer $UNIFY_KEY"
```

--------------------------------

### POST /v0/chat/completions - Anthropic Example with Beta Headers

Source: https://docs.unify.ai/universal_api/queries

Shows how to use provider-specific beta features via custom headers, such as `anthropic-beta`, when interacting with models through the Unify API.

```APIDOC
## POST /v0/chat/completions (Beta Headers)

### Description
Sends a chat completion request with custom headers to enable beta features for specific providers, like Anthropic.

### Method
POST

### Endpoint
/v0/chat/completions

### Parameters
#### Request Body
- **model** (string) - Required - The model to use (e.g., "claude-3.5-sonnet@anthropic").
- **messages** (array) - Required - An array of message objects.

#### Request Headers
- **anthropic-beta** (string) - Required - Header to enable specific beta features (e.g., "max-tokens-3-5-sonnet-2024-07-15").

### Request Example
```json
{
    "model": "claude-3.5-sonnet@anthropic",
    "messages": [
        {
            "content": "Tell me a joke",
            "role": "user"
        }
    ]
}
```

### Response
#### Success Response (200)
- **content** (string) - The generated response from the model.

#### Response Example
```json
{
  "content": "Why did the scarecrow win an award? Because he was outstanding in his field!"
}
```
```

--------------------------------

### Create Assistant Request - Python

Source: https://docs.unify.ai/api-reference/assistant_management/create_a_new_assistant

Example of how to create a new AI assistant using Python's requests library. This demonstrates sending a POST request with a JSON payload containing assistant configuration details. It requires an API token for authentication.

```python
import requests

url = "https://api.unify.ai/v0/assistant"

headers = {"Authorization": "Bearer <token>"}

json_input = {"first_name": "Ada", "surname": "Lovelace", "age": 28, "weekly_limit": 15.75, "max_parallel": 2, "nationality": "North America", "profile_photo": "https://example.com/photos/ada.jpg", "profile_video": "https://example.com/videos/ada.mp4", "desktop_url": "https://app.example.com/assistants/ada", "user_local_desktop": "windows", "about": "Mathematician and writer known for work on Analytical Engine", "email": "ada.lovelace@unify.ai", "voice_id": "bf0a246a-8642-498a-9950-80c35e9276b5", "voice_provider": "elevenlabs", "voice_mode": "tts", "user_phone": "+15551234567", "user_whatsapp_number": "+15551234567", "phone_country": "US", "timezone": "America/New_York"}

response = requests.request("POST", url, json=json_input, headers=headers)

print(response.text)
```

--------------------------------

### GET /v0/metrics

Source: https://docs.unify.ai/api-reference/usage/get_query_metrics

Fetches aggregated telemetry data from previous queries to the /chat/completions endpoint.

```APIDOC
## GET /v0/metrics

### Description
Returns aggregated telemetry data from previous queries to the `/chat/completions` endpoint, specifically the p50 and p95 for generation time and tokens per second, and also the total prompt and completion tokens processed within the interval. The user id and total request count within the interval are also returned.

### Method
GET

### Endpoint
`/v0/metrics`

### Parameters
#### Header Parameters
- **Authorization** (string) - Required - Bearer authentication header of the form `Bearer <token>`, where `<token>` is your auth token.

#### Query Parameters
- **start_time** (string) - Optional - Timestamp of the earliest query to aggregate. Format is `YYYY-MM-DD hh:mm:ss`.
- **end_time** (string) - Optional - Timestamp of the latest query to aggregate. Format is `YYYY-MM-DD hh:mm:ss`.
- **models** (string) - Optional - Models to fetch metrics from. The list must be a set of comma-separated strings. i.e. `gpt-3.5-turbo,gpt-4o`.
- **providers** (string) - Optional - Providers to fetch metrics from. The list must be a set of comma-separated strings. i.e. `openai,together-ai`.
- **interval** (string) - Optional - Number of seconds in the aggregation interval. Defaults to `300`.
- **secondary_user_id** (string) - Optional - Secondary user id. The secondary user id will match any string previously sent in the `user` attribute of `/chat/completions`.

### Request Example
```bash
curl --request GET \
  --url 'https://api.unify.ai/v0/metrics?start_time=2024-07-12%2004:20:32&end_time=2024-08-12%2004:20:32&models=gpt-4o,llama-3.1-405b-chat,claude-3.5-sonnet&providers=openai,anthropic,fireworks-ai&interval=300&secondary_user_id=sample_user_id' \
  --header "Authorization: Bearer $UNIFY_KEY"
```

```python
import requests

url = "https://api.unify.ai/v0/metrics?start_time=2024-07-12%2004:20:32&end_time=2024-08-12%2004:20:32&models=gpt-4o,llama-3.1-405b-chat,claude-3.5-sonnet&providers=openai,anthropic,fireworks-ai&interval=300&secondary_user_id=sample_user_id"

headers = {"Authorization": "Bearer <token>"}

response = requests.request("GET", url, headers=headers)

print(response.text)
```

### Response
#### Success Response (200)
(Response body is empty for success)

#### Error Response (422)
- **detail** (array) - Details about the validation error.
  - **loc** (array) - Location of the error in the request.
  - **msg** (string) - Error message.
  - **type** (string) - Error type.

#### Response Example
```json
{
    "detail": [
        {
            "loc": ["string"],
            "msg": "string",
            "type": "string"
        }
    ]
}
```
```

--------------------------------

### Client Configuration Methods

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/AsyncMultiUnify

Methods for configuring the client's default behavior, such as setting system messages, tags, temperature, tool choices, and more.

```APIDOC
## Client Configuration Methods

### set_system_message

#### Description
Set the default system message for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (str) - Required - The default system message.

#### Returns
This client, useful for chaining inplace calls.

### set_tags

#### Description
Set the default tags for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (List[str]) - Required - The default tags.

#### Returns
This client, useful for chaining inplace calls.

### set_temperature

#### Description
Set the default temperature for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (float) - Required - The default temperature.

#### Returns
This client, useful for chaining inplace calls.

### set_tool_choice

#### Description
Set the default tool choice for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (ChatCompletionToolChoiceOptionParam) - Required - The default tool choice.

#### Returns
This client, useful for chaining inplace calls.

### set_tools

#### Description
Set the default tools for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (Iterable[ChatCompletionToolParam]) - Required - The default tools.

#### Returns
This client, useful for chaining inplace calls.

### set_top_logprobs

#### Description
Set the default top logprobs for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (int) - Required - The default top logprobs.

#### Returns
This client, useful for chaining inplace calls.

### set_top_p

#### Description
Set the default top p value for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (float) - Required - The default top p value.

#### Returns
This client, useful for chaining inplace calls.

### set_traced

#### Description
Set the default traced bool for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (bool) - Required - The default traced bool.

#### Returns
This client, useful for chaining inplace calls.

### set_use_custom_keys

#### Description
Set the default use custom keys bool for the client.

#### Method
This is a client method, not a direct HTTP endpoint.

#### Parameters
* `value` (bool) - Required - The default use custom keys bool.

#### Returns
This client, useful for chaining inplace calls.

```

--------------------------------

### Manage Output Explanations for Sub-questions

Source: https://docs.unify.ai/case_study/optimize_agent/iteration_3

Defines a dictionary `output_response_explanations` containing two distinct explanations for populating the output fields: one for scenarios with sub-questions and another for scenarios without. This allows the system prompt to guide the agent on how to format its response based on the question's structure.

```python
output_response_explanations = dict()
output_response_explanations["with_subqs"] = (
    "For each sub-question {subquestions} you should populate the `reasoning` field with your initial reasoning about the correct number of marks to award. Finally, you should put the number of marks to award for this sub-question in the `marks` field."
)
output_response_explanations["without_subqs"] = (
    "You should populate the `reasoning` field with your initial reasoning about the correct number of marks to award. Finally, you should put the number of marks to award in the `marks` field."
)
```

--------------------------------

### Get Query Tags

Source: https://docs.unify.ai/api-reference/usage/get_query_tags

Retrieves a list of query tags. Requires Bearer authentication.

```APIDOC
## GET /v0/tags

### Description
Retrieves a list of query tags. Requires Bearer authentication.

### Method
GET

### Endpoint
/v0/tags

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Request Example
```bash cURL
curl --request GET \
  --url 'https://api.unify.ai/v0/tags' \
  --header "Authorization: Bearer $UNIFY_KEY"
```

```python Python
import requests

url = "https://api.unify.ai/v0/tags"

headers = {"Authorization": "Bearer <token>"}

response = requests.request("GET", url, headers=headers)

print(response.text)
```

### Response
#### Success Response (200)
- **None** (object) - An empty object is returned upon successful retrieval of tags.

#### Response Example
```json 200
{}
```
```

--------------------------------

### GET /v0/assistant/photo/animate/{prediction_id}

Source: https://docs.unify.ai/api-reference/media/get_animation_prediction_status

Retrieves the status and result of a video animation job using the provided prediction ID.

```APIDOC
## GET /v0/assistant/photo/animate/{prediction_id}

### Description
Retrieves the status and result of a video animation job.

### Method
GET

### Endpoint
`/v0/assistant/photo/animate/<prediction_id>`

### Parameters
#### Path Parameters
- **prediction_id** (string) - Required - The ID of the animation prediction job.

#### Query Parameters
None

#### Request Body
None

### Request Example
```bash
curl --request GET \
  --url 'https://api.unify.ai/v0/assistant/photo/animate/<prediction_id>' \
  --header "Authorization: Bearer $UNIFY_KEY"
```

```python
import requests

url = "https://api.unify.ai/v0/assistant/photo/animate/<prediction_id>"

headers = {"Authorization": "Bearer <token>"}

response = requests.request("GET", url, headers=headers)

print(response.text)
```

### Response
#### Success Response (200)
- **info** (object) - Contains information about the animation job status and result.

#### Error Response (422)
- **detail** (array) - An array of error details, each containing `loc`, `msg`, and `type`.

#### Response Example
```json
{
    "info": "object"
}
```

```json
{
    "detail": [
        {
            "loc": "string",
            "msg": "string",
            "type": "string"
        }
    ]
}
```
```

--------------------------------

### Create Assistant Request - cURL

Source: https://docs.unify.ai/api-reference/assistant_management/create_a_new_assistant

Example of how to create a new AI assistant using a cURL command. This request includes all optional and required parameters for configuring the assistant's profile, limits, and voice settings. It requires an authorization token for authentication.

```bash
curl --request POST \
  --url 'https://api.unify.ai/v0/assistant' \
  --header "Authorization: Bearer $UNIFY_KEY" \
  --header 'Content-Type: application/json' \
  --data '{ \
    "first_name": "Ada", \
    "surname": "Lovelace", \
    "age": 28, \
    "weekly_limit": 15.75, \
    "max_parallel": 2, \
    "nationality": "North America", \
    "profile_photo": "https://example.com/photos/ada.jpg", \
    "profile_video": "https://example.com/videos/ada.mp4", \
    "desktop_url": "https://app.example.com/assistants/ada", \
    "user_local_desktop": "windows", \
    "about": "Mathematician and writer known for work on Analytical Engine", \
    "email": "ada.lovelace@unify.ai", \
    "voice_id": "bf0a246a-8642-498a-9950-80c35e9276b5", \
    "voice_provider": "elevenlabs", \
    "voice_mode": "tts", \
    "user_phone": "+15551234567", \
    "user_whatsapp_number": "+15551234567", \
    "phone_country": "US", \
    "timezone": "America/New_York" \
}'
```

--------------------------------

### GET /v0/queries

Source: https://docs.unify.ai/api-reference/usage/get_queries

Retrieves a list of queries with options to filter by tags, endpoints, time range, page number, and failure status.

```APIDOC
## GET /v0/queries

### Description
Retrieves a list of queries with options to filter by tags, endpoints, time range, page number, and failure status.

### Method
GET

### Endpoint
/v0/queries

### Parameters
#### Header Parameters
- **Authorization** (string) - Required - Bearer authentication header of the form `Bearer <token>`.

#### Query Parameters
- **tags** (string | array | null) - Optional - Tags to filter for queries that are marked with these tags.
- **endpoints** (string | array | null) - Optional - Optionally specify an endpoint, or a list of endpoints to filter for.
- **start_time** (string | null) - Optional - Timestamp of the earliest query to aggregate. Format is `YYYY-MM-DD hh:mm:ss`.
- **end_time** (string | null) - Optional - Timestamp of the latest query to aggregate. Format is `YYYY-MM-DD hh:mm:ss`.
- **page_number** (integer | null) - Optional - The query history is returned in pages, with up to 20 prompts per page. Defaults to 1.
- **failures** (boolean | string) - Optional - Indicates whether to include failures in the return (when set as True), or whether to return failures exclusively (when set as 'only'). Defaults to False.

### Request Example
```bash
curl --request GET \
  --url 'https://api.unify.ai/v0/queries?tags=my_tag&endpoints=gpt-4o@openai&start_time=2024-07-12%2004:20:32&end_time=2024-08-12%2004:20:32&page_number=1&failures=False' \
  --header "Authorization: Bearer $UNIFY_KEY"
```

```python
import requests

url = "https://api.unify.ai/v0/queries?tags=my_tag&endpoints=gpt-4o@openai&start_time=2024-07-12%2004:20:32&end_time=2024-08-12%2004:20:32&page_number=1&failures=False"

headers = {"Authorization": "Bearer <token>"}

response = requests.request("GET", url, headers=headers)

print(response.text)
```

### Response
#### Success Response (200)
(No specific fields documented for success response, an empty JSON object is shown as an example)

#### Response Example
```json
{}
```

#### Error Response (422)
- **detail** (array) - Array of error details.
  - **loc** (array) - Location of the error.
  - **msg** (string) - Error message.
  - **type** (string) - Error type.

#### Error Response Example
```json
{
    "detail": [
        {
            "loc": [
                "string"
            ],
            "msg": "string",
            "type": "string"
        }
    ]
}
```
```

--------------------------------

### Install Tracing Hook (Python)

Source: https://docs.unify.ai/python/functions/logging/utils/tracing/install_tracing_hook

Installs an import hook to wrap specified modules with a tracing decorator. It adds a TraceFinder to sys.meta_path to intercept module imports. The hook is only installed if it doesn't already exist. Arguments include a list of target module name prefixes and an optional filter function.

```python
from typing import List, Callable

def install_tracing_hook(targets: List[str], filter: Callable = None):
    # Function implementation details would follow here
    pass
```

--------------------------------

### Run Experiment for Evaluating Queries Per Mark in Python

Source: https://docs.unify.ai/case_study/optimize_agent/iteration_9

This Python snippet sets up and runs an experiment named 'queries_per_mark' using the `unify` library. It configures parameters including system messages, dataset, and source, then maps the `evaluate` function to a test dataset. This is used to test the evaluation logic with different system message configurations.

```python
with unify.Experiment(
    "queries_per_mark",
    overwrite=True,
), unify.Params(
    subq_system_message=subq_system_message,
    mark_system_message=mark_system_message,
    dataset="TestSet10",
    source=unify.get_source(),
):
    unify.map(
        evaluate,
        [
            dict(
                **d.entries,
                _subq_system_message=subq_system_message,
                _mark_system_message=mark_system_message,
            )
            for d in test_set_10
        ],
        name="Evals",
    )
```

--------------------------------

### Initialize AsyncMultiUnify Client (Python)

Source: https://docs.unify.ai/python/classes/universal_api/clients/multi_llm/AsyncMultiUnify

Initializes the AsyncMultiUnify client with various configuration options for interacting with LLM endpoints. It supports specifying endpoints, system messages, conversation history, and fine-tuning generation parameters like temperature and top_p. It also includes arguments for platform-specific settings, client behavior, and passthrough parameters for requests.

```python
class AsyncMultiUnify:
    def __init__(
        self,
        endpoints: Optional[Union[str, Iterable[str]]] = None,
        *, 
        system_message: Optional[str] = None,
        messages: Optional[
            Union[
                List[ChatCompletionMessageParam],
                Dict[str, List[ChatCompletionMessageParam]],
            ]
        ] = None,
        frequency_penalty: Optional[float] = None,
        logit_bias: Optional[Dict[str, int]] = None,
        logprobs: Optional[bool] = None,
        top_logprobs: Optional[int] = None,
        max_completion_tokens: Optional[int] = None,
        n: Optional[int] = None,
        presence_penalty: Optional[float] = None,
        response_format: Optional[Union[Type[BaseModel], Dict[str, str]]] = None,
        seed: Optional[int] = None,
        stop: Union[Optional[str], List[str]] = None,
        temperature: Optional[float] = 1.0,
        top_p: Optional[float] = None,
        service_tier: Optional[str] = None,
        tools: Optional[Iterable[ChatCompletionToolParam]] = None,
        tool_choice: Optional[ChatCompletionToolChoiceOptionParam] = None,
        parallel_tool_calls: Optional[bool] = None,
        reasoning_effort: Optional[str] = None,
        # platform arguments
        use_custom_keys: bool = False,
        tags: Optional[List[str]] = None,
        drop_params: Optional[bool] = True,
        region: Optional[str] = None,
        log_query_body: Optional[bool] = True,
        log_response_body: Optional[bool] = True,
        api_key: Optional[str] = None,
        # python client arguments
        stateful: bool = False,
        return_full_completion: bool = False,
        traced: bool = False,
        direct_mode: bool = False,
        cache: Union[bool, str] = None,
        cache_backend: Optional[str] = None,
        # passthrough arguments
        extra_headers: Optional[Headers] = None,
        extra_query: Optional[Query] = None,
        **kwargs,
    ) -> None: 
        pass
```

--------------------------------

### Build System Message with Dynamic Output Explanation

Source: https://docs.unify.ai/case_study/optimize_agent/iteration_3

Constructs a system message for an AI agent, dynamically incorporating guidelines, question details, markscheme, student's answer, and a specific explanation for output formatting. It handles conditional logic for including sub-question details in the output explanation.

```python
system_message = """
Your task is to award a suitable number of marks for a student's answer to a question, from 0 up to a maximum of {available_marks_total} marks.

The general marking guidelines (relevant for all questions) are as follows:

{general_guidelines}

The question you need to mark is:

{question}


The markscheme for this specific question is:

{markscheme}


The student's answer to this question (which you need to marked) is:

{answer}


{output_response_explanation}
""".replace(
    "{general_guidelines}",
    general_guidelines,
)
```

--------------------------------

### Get Animation Prediction Status using cURL

Source: https://docs.unify.ai/api-reference/media/get_animation_prediction_status

Fetches the status of a video animation prediction job using a cURL command. It requires the prediction ID as a path parameter and an Authorization header with a Bearer token.

```bash
curl --request GET \
  --url 'https://api.unify.ai/v0/assistant/photo/animate/<prediction_id>' \
  --header "Authorization: Bearer $UNIFY_KEY"
```

--------------------------------

### Get Credit Balance via HTTP Request

Source: https://docs.unify.ai/basics/billing

This snippet demonstrates how to retrieve your current credit balance using a GET HTTP request to the Unify AI API. It requires an authorization token and specifies the expected JSON response format.

```shell
curl -X 'GET' \
    'https://api.unify.ai/v0/credits' \
    -H 'accept: application/json' \
    -H "Authorization: Bearer $UNIFY_KEY"
```

```json
{
    "id": "your_user_id",
    "credits": 232.32
}
```

--------------------------------

### Get Animation Prediction Status using Python

Source: https://docs.unify.ai/api-reference/media/get_animation_prediction_status

Retrieves the status of a video animation prediction job using Python's requests library. This method requires the prediction ID and an Authorization header containing the Bearer token.

```python
import requests

url = "https://api.unify.ai/v0/assistant/photo/animate/<prediction_id>"

headers = {"Authorization": "Bearer <token>"}

response = requests.request("GET", url, headers=headers)

print(response.text)
```
Tokens:
Raw