You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
docs(readme): reframe Scope table — drop empty compute row, 'Provided by your stack'
Renames the gap column to a reuse-your-stack framing and folds the cron caveat into the orchestration row (durable step execution, not wall-clock triggering).
Copy file name to clipboardExpand all lines: README.md
+3-4Lines changed: 3 additions & 4 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -35,15 +35,14 @@ Atmosphere is built for teams that need AI agents to behave like production serv
35
35
36
36
Atmosphere is a JVM framework, not an agent-hosting platform. We ship the primitives your application uses at runtime; the host you choose (Tomcat, Jetty, Netty, Undertow, Quarkus, Spring Boot, or any servlet container) owns compute and scheduling. Compared to the agent-platform stack vocabulary that has emerged around offerings like Cloudflare Agents, AWS Bedrock Agents, and Vertex AI Agents:
37
37
38
-
| Layer | What Atmosphere ships |What Atmosphere does not ship|
38
+
| Layer | What Atmosphere ships |Provided by your stack|
39
39
|---|---|---|
40
-
| Compute / scheduling | — | Your JVM host or container scheduler |
41
40
| Streaming transport |`atmosphere-runtime` over WebTransport/HTTP3, WebSocket, SSE, long-polling, gRPC | — |
42
41
| Runtime dispatch |`atmosphere-ai``AgentRuntime` SPI + 12 adapters with contract-tested capability flags | Model hosting (we call providers; we do not host weights) |
43
-
| Orchestration |`@Coordinator`, `AgentFleet`, handoffs, conditional routing, event-sourced coordination journal (`CoordinationJournal` SPI with causal `EventEnvelope` lineage, `CoordinationProjection` DAG-from-log, `FileCoordinationJournal` append-only NDJSON persistence, `CoordinationFork` what-if branching), result evaluation, and durable hibernating `Workflow<S>` over `CheckpointStore` (per-step retry, resume across JVM restart, no thread held while hibernated) | Long-running cron / scheduled execution (use a dedicated scheduler) |
42
+
| Orchestration |`@Coordinator`, `AgentFleet`, handoffs, conditional routing, event-sourced coordination journal (`CoordinationJournal` SPI with causal `EventEnvelope` lineage, `CoordinationProjection` DAG-from-log, `FileCoordinationJournal` append-only NDJSON persistence, `CoordinationFork` what-if branching), result evaluation, and durable hibernating `Workflow<S>` over `CheckpointStore` (per-step retry, resume across JVM restart, no thread held while hibernated) — durable step execution, not wall-clock triggering | Cron / wall-clock scheduling (your container scheduler or a dedicated scheduler fires the workflow) |
44
43
| Memory |`AiConversationMemory` per-conversation history (in-memory, plus durable SQLite/Redis through the `ConversationPersistence` SPI in `atmosphere-durable-sessions{-sqlite,-redis}`), `LongTermMemory` per-user facts (`InMemoryLongTermMemory`, `SqliteLongTermMemory`, `RedisLongTermMemory`), `SemanticRecallInterceptor` for BYO vector-store recall | Managed vector stores (use Spring AI's `VectorStore`, LangChain4j embeddings, or your own) |
0 commit comments