Workshop Part 1

get a working unified interface in this section, you will connect a provider add models with stable aliases create a workload generate a secure workload key send your first request by the end, you will have real traffic flowing through the amberflo unified interface create your account go to amberflo ai and click start free trial complete the signup form, confirm your email, and log in you will land on the dashboard with demo mode enabled start from a clean state turn off demo mode navigate to ai spend the page should be empty that is expected nothing has been configured yet you are now starting from zero connect a provider providers allow the amberflo unified interface to route traffic to ai platforms such as openai or aws bedrock in the providers section click setup provider select your provider (for example, openai) paste in your api key name the credentialexample workshop openai save amberflo can now send traffic to this provider add models applications reference models by alias , not by provider specific names alias rules all lowercase no spaces use hyphens add a base model in models click add model select the credential you just created choose a model (for example, gpt 4o) set the alias simple model test the connection save add a second model add another model using the same credential for example underlying model gpt 5 2 alias latest model you now have multiple models available through a single interface why aliases matter when a new model version is released, you can update the underlying provider model without changing application code the alias remains stable define a workload workloads represent applications or systems that consume ai in workloads click add workload name it customer support bot leave the generated id unchanged select the models simple model latest model save this workload defines which models the application can access the boundary for usage and cost attribution\\ generate a workload key without a workload key, no traffic can flow in workload keys click add workload key select customer support bot name the key customer support bot key save a modal will display the key value ⚠️ copy it and store it securely you will not be able to retrieve it later send your first request open the model playground paste in your workload key select simple model enter a prompt, for example\ what is an llm? submit if you receive a response, the amberflo unified interface is fully configured and working what you now have you have a provider connected multiple models configured a workload defined a secure workload key a successful request flowing through amberflo you now have a functioning amberflo unified interface next integrate it into your application code