A prompt engineering solution to manage Gen AI prompts easily.
- Self-hosted option with full control over data
- Dockerized for easy deployment
- RESTful API for easy integration
- With SDK for Python and Node.js.
- API Key management through centralized UI
- Prompt Management through centralized UI
- Variants
- Versioning (database level)
- PromptSmith Intro This is a video to demonstrate the features of Prompt Smith
- Start the service locally
docker-compose up --build
- Access the at
http://localhost:3000
- Default login credentials (this could be set in docker compose file, please do not use default credentials in production)
- email:
admin@localhost.lan
- password:
AwesomePromptsManagement
- email:
- Create a prompt at
http://localhost:3000/prompt
- Create an API key at
http://localhost:3000/key
- Get your prompts with ease through RESTful API
curl "http://localhost:3000/api/bff/api/sdk/prompt/{prompt-key}" \
-H 'Authorization: Bearer {api-key}' \
-H 'Content-Type: application/json; charset=utf-8'
- Or Get prompts through SDK
We have deployed a Demo Instance
- Please note the data may not be persisted and will be deleted scheduledly.
These features were not planned in the beginning but could be added in the future.
- Prompt Compression, to reduce the cost of token usage but still maintain the quality of prompts
- SaaS version of Prompt Smith.
- Trace of prompt usage.
- Webhook support.
- Prompt Management through CLI.
- Integration with analytics tools, such as Google Analytics, Mixpanel, etc.
Please feel free to open an issue or PR for any feature requests or bug reports.