conproxy 0.2.1

Cache proxy server for heterogeneous RAG/vector search backends with cascade, scope, and MCP
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

# Used by Tilt with K8s + kind.
# llama-server runs on host (not in pod); reachable via host.docker.internal.
apiVersion: v1
kind: ConfigMap
metadata:
  labels:
    app: conproxy
data:
  conproxy.toml: |
    id = "llama"
    provider = "openai"
    base_url = "http://host.docker.internal:8081"
    api_key = "sk-local"
    timeout_secs = 120

    default_timeout_secs = 120
    cache_enabled = true
    cache_ttl_secs = 120
    max_body_bytes = 4194304
    max_cache_entries = 500