Stop rewriting cloud infrastructure for every AI project. Launch immediately on our Hosted AI Agent Cloud Studio or deploy to your own AWS & GCP accounts with real-time SSE token streaming, DynamoDB & Firestore session memory, S3 & GCS presigned uploads, and async background queues.
POST /v1/chat/stream & POST /v1/jobs
Click "▶ Stream SSE (/v1/chat/stream)" or choose a sample prompt above to stream tokens from the FastAPI ResearchAgent...
Our unified cloud adapter automatically routes conversation memory, job progress, presigned file uploads, and background queues to native AWS or Google Cloud serverless services.
| Layer / Capability | Amazon Web Services (AWS Serverless) | Google Cloud Platform (Cloud Run) |
|---|---|---|
| Serverless Compute & SSE Streaming | AWS Lambda Function URL / ECS Fargate | Google Cloud Run (HTTP/2 + SSE up to 60m) |
Agent Memory (SessionState) |
Amazon DynamoDB (PK=SESSION#<id> + native TTL) |
Cloud Firestore (agent_sessions + native TTL) |
Async Job Tracking (JobRecord) |
Amazon DynamoDB (PK=JOB#<id> atomic list_append) |
Cloud Firestore (agent_jobs document updates) |
| Object Storage & Presigned URLs | Amazon S3 + V4 Pre-signed PUT/GET URLs | Google Cloud Storage + IAM signBlob V4 URLs |
| Background Job Queue | Amazon SQS + DLQ → Worker Lambda trigger | Google Cloud Tasks → OIDC push to /v1/worker |
| 1-Command Infrastructure | make deploy-aws (AWS CDK Python) | make deploy-gcp (Cloud Run + Terraform) |
Plan → Tool Execution → Synthesis supporting Google Gemini & OpenAIboto3 & google-cloud-*/dashboard): Create custom agents, attach S3/GCS files, and manage API keysdeploy_cloudrun.sh + declarative TerraformStart free in our Cloud Studio or upgrade to Pro for 5,000 monthly agent runs & full commercial deployment rights.
5,000 hosted Pro agent runs/month + custom agents + AWS CDK & GCP Cloud Run deployment kit.
deploy_aws.sh)deploy_cloudrun.sh)