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
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

# Federated search: local Qdrant + remote Elasticsearch.
# Requires: both backends running.
#
# Two legs on contexts.default; federated policy is per-context.

[server]
listen = "127.0.0.1:9090"

[upstreams.local-qdrant]
url = "http://localhost:6333"
type = "qdrant"

[upstreams.remote-es]
url = "http://localhost:9200"
type = "elasticsearch"

[contexts.default]
default = true

[[contexts.default.upstreams]]
ref = "local-qdrant"
priority = 0

[[contexts.default.upstreams]]
ref = "remote-es"
priority = 1

[contexts.default.cache]
fresh_secs = 300
stale_secs = 600
max_entries = 10000

[contexts.default.federated]
enabled = true
min_local_results = 3
min_local_confidence = 0.5
fallback_on_empty = true
fallback_on_low_confidence = true
merge_mode = "local_priority"
max_merged_results = 20