Multimodal AI you can deploy anywhere
Reka is an AI research and product company that develops multimodal models that understand the world, reason, and take actions.
Built to work with multimodal streams of video, audio, and text.
Reka Vision Agent
Search across millions of videos or images using natural language, summarize long videos, detect complex events to trigger live alerts.
Reka Speech Agent
Transcribe and translate in real time for speech-to-text or speech-to-speech applications.
Reka Research Agent
Browse the web and internal file systems to answer research questions from multiple sources.
Accessed via Simple and Powerful APIs
url = f"{BASE_URL}/videos/upload" # Video upload endpoint
video_path = "/content/demo.mp4" # Sample video for demo purposes
# Video indexing request body
data = {
"user_id": "test_user", # id of user this video will be stored under
"index": True, # whether the video should be indexed for search/qa/etc
"video_name": "vid16", # name of video to store in vision agent system
}
headers = {
"X-Api-Key": REKA_API_KEY
}
# Open the video file and send the request
with open(video_path, "rb") as file:
files = {"file": (video_path[1:], file, "video/mp4")} # Send as multipart/form-data
response = requests.post(url, headers=headers, data=data, files=files)
# Print response
if response.status_code != 500:
print(response.status_code, response.json())
else:
print(response.status_code, response.text)
from reka import ChatMessage
from reka.client import Reka
client = Reka()
response = client.chat.create(
messages=[
ChatMessage(
content=[
{"type": "image_url", "image_url": "https://v0.docs.reka.ai/_images/000000245576.jpg"},
{"type": "text", "text": "What animal is this? Answer briefly"}
],
role="user",
)
],
model="reka-core-20240501",
)
print(response.responses[0].message.content)
