Skip to content

The cross-harness, self-improving memory layer for AI agents. Join our community: https://discord.com/invite/zdNChS2fBu

README: mention LLM gateway support

Claudefb8eae818d8a
diff --git a/README.md b/README.md
index fb00d2af..f4b378bc 100644
--- a/README.md
+++ b/README.md
@@ -379,6 +379,12 @@ The [Beacon Skills page](https://docs.beacon.sh/concepts/beacon-skills#install)
| Claude Agent SDK | `Observe.wrapClaudeAgentQuery()` | Query root spans with Beacon-compatible prompt attributes |
| Vercel AI SDK | `experimental_telemetry` tracer handoff | Model call and tool spans where telemetry is enabled |
+### LLM Gateways
+
+| Gateway | Collection | Captures |
+| --- | --- | --- |
+| LiteLLM, Envoy (incl. Envoy Gateway, Istio), Portkey | [`beacon-gateway`](gateway/README.md) hook adapter | Prompts, responses, tool calls, and token usage for every model call routed through the gateway |
+
---
## Output Destinations
@@ -417,7 +423,8 @@ Local agents ───────┐
Browser chat ───────┤
CI ─────────────────┼──→ Beacon ──→ unified session history
Cloud agents ───────┤ │
-Agent SDKs ─────────┘ ├──→ local JSONL
+Agent SDKs ─────────┤ ├──→ local JSONL
+LLM gateways ───────┘ │
├──→ reviewed memory
├──→ MCP / Agent Skills
└──→ your own infrastructure