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
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55

# Two contexts share the same Meilisearch resource with different
# index/timeout overrides + independent caches and scope phrases.
#
# Requires: Meilisearch on :7700 (see meilisearch-quickstart.toml).

[server]
listen = "127.0.0.1:9090"

[upstreams.meili]
url = "http://localhost:7700"
type = "meilisearch"
api_key = "dev_master_key"
timeout_secs = 30

[contexts.docs]
default = true
description = "Product docs"

[[contexts.docs.upstreams]]
ref = "meili"
index = "docs"
timeout_secs = 5

[contexts.docs.cache]
fresh_secs = 300
stale_secs = 3600
max_entries = 10000

[contexts.docs.scope]
mode = "filter"
min_similarity = 0.25

[[contexts.docs.scope.weighted_phrases]]
text = "product api"
weight = 1.0

[contexts.support]
description = "Support tickets"

[[contexts.support.upstreams]]
ref = "meili"
index = "support"
timeout_secs = 15

[contexts.support.cache]
fresh_secs = 60
max_entries = 5000

[contexts.support.scope]
mode = "filter"
min_similarity = 0.35

[[contexts.support.scope.weighted_phrases]]
text = "billing"
weight = 1.2