Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/cognosisai/platform
https://github.com/cognosisai/platform
ai docker elasticsearch gpt-3 javascript llm typescript
Last synced: about 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/cognosisai/platform
- Owner: cognosisai
- License: mit
- Created: 2022-10-01T19:04:07.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2023-02-04T23:30:18.000Z (over 1 year ago)
- Last Synced: 2024-03-14T20:04:51.482Z (6 months ago)
- Topics: ai, docker, elasticsearch, gpt-3, javascript, llm, typescript
- Language: TypeScript
- Homepage:
- Size: 351 KB
- Stars: 168
- Watchers: 6
- Forks: 6
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-temporal - Cognosis AI Platform - Template for large language model applications. (TypeScript / Libraries)
- awesome-ChatGPT-repositories - platform - @erikschluntz when i initially played with gpt-3 agents in the summer of last year (you can see an example here: whereby you can give a gpt-3 agent with access to a javascript repl a goal to accomplish), that's where i ended up, and i didn't think it was that… (NLP)
README
# Cognosis AI Platform
## Summary
The Cognosis AI Platform contains an application server and all of the
infrastructure you need to build Large Language Model applications with,
batteries included!
##
[Join our Slack!](https://slack.com/get-started?sig=enQtNDczNzkwODAxODM4NC1iMzAzNTA4MmE1Zjk1NDQyMTk5ZmZhMjFiMzMyMzJjNGMxZGUyMDBjYjA4YjcxYjJjOTRmNjRlNmY4MzU2OTU0&invite_id=I04MPSQ0JBA&redir=%2Fshare%2FenQtNDczNzkwODAxODM4NC1iMzAzNTA4MmE1Zjk1NDQyMTk5ZmZhMjFiMzMyMzJjNGMxZGUyMDBjYjA4YjcxYjJjOTRmNjRlNmY4MzU2OTU0#/sharedinvite)
## Quickstart
Requirements:
* Docker
* Node 19
### Step 1: Get Infrastructure Running
```
# Clone the repo and cd into it
git clone [email protected]:cognosisai/platform.git
cd platform
# On Apple M1/M2 chips:
make build-apple
# On x86:
make build-x86
# Spin it all up! This runs docker-compose up, and will get you
# Elasticsearch 8.5, cognosis-embeddings service, and temporalite, which
# is a single Docker container version of Temporal meant for development
make run
```
### Step 2: Configure
Edit .env, and populate it with the following information:
```
TEMPORAL_HOST="localhost:7233"
ELASTIC_CONFIG={"auth":{"username":"elastic","password":"changeme"},"node":"http://localhost:9200"}
OPENAI_TOKEN=""
NLPCLOUD_TOKEN=""
EMBEDDINGS_URL="http://localhost:9100"
```
### Step 3: Install NPM modules
```
npm install
```
## Step 4: Start Temporal Worker
```
ts-node src/worker.ts
```
This will take a few seconds. It will be ready when you see this:
```
2022-11-05T23:51:17.087Z [INFO] webpack 5.74.0 compiled successfully in 717 ms
2022-11-05T23:51:17.093Z [INFO] Workflow bundle created { size: '0.91MB' }
2022-11-05T23:51:17.141Z [INFO] Worker state changed { state: 'RUNNING' }
```
## Step 5: AI!
```
ts-node src/cli/cognosis.ts "Give me a really terrible idea involving an \
icepick, and bottle of elmer's wood glue"
Starting wfid workflow-fC-ONstofj4T4G9C_JQv5
Use the icepick to make a hole in the bottle of glue, then drink it.
```
### Application Server Components
#### Elastic Search
#### Temporal
#### Embeddings (Tensorflow - Google USE5)
#### Cognosis AI SDK
### Cognosis AI SDK
Cognosis AI Platform includes