docs(config): rename sample [router] sections to [talkie]

This commit is contained in:
2026-07-03 13:26:48 +04:00
parent 595ec187bc
commit 0c8a7e88f6
+10 -9
View File
@@ -75,7 +75,7 @@ capabilities = { General = 1.0, Coding = 0.7, Reasoning = 0.6, Summarization = 0
# Project-scoped, cross-session memory (distilled decision journal + repo map). # Project-scoped, cross-session memory (distilled decision journal + repo map).
# When enabled, decisions are distilled to durable per-repo memory at session end and # When enabled, decisions are distilled to durable per-repo memory at session end and
# retrieved at the next session's start via the router L3 path. # retrieved at the next session's start via the L3 path.
[project] [project]
enabled = false enabled = false
root = "" # repo root key; empty = current working dir root = "" # repo root key; empty = current working dir
@@ -86,29 +86,30 @@ max_depth = 4 # directory recursion cap
inject_top_k = 30 # entries injected into context (ranked by score) inject_top_k = 30 # entries injected into context (ranked by score)
# ignore_globs = [".git", "node_modules", "build", "target", ".gradle", "dist", ".idea"] # ignore_globs = [".git", "node_modules", "build", "target", ".gradle", "dist", ".idea"]
# Router configuration (optional, defaults shown below) # Talkie: the conversational front-end (CHAT triage + STEERING into a running workflow).
[router] # (Legacy section name [router] is still read as a fallback.)
[talkie]
conversation_keep_last = 6 # how many recent chat turns to keep in context conversation_keep_last = 6 # how many recent chat turns to keep in context
retrieval_k = 5 # L3 memory hits to retrieve per query retrieval_k = 5 # L3 memory hits to retrieve per query
token_budget = 4096 # context budget for router prompts token_budget = 4096 # context budget for Talkie prompts
# Sampling/length for router chat + steering replies. # Sampling/length for Talkie chat + steering replies.
[router.generation] [talkie.generation]
temperature = 0.7 temperature = 0.7
top_p = 0.9 top_p = 0.9
max_tokens = 512 max_tokens = 512
# Router narration (the live workflow commentary). max_tokens is higher than chat so # Narration (the live workflow commentary). max_tokens is higher than chat so
# reasoning models can finish "thinking" and still emit the line; max_per_run caps how # reasoning models can finish "thinking" and still emit the line; max_per_run caps how
# many narrations fire per workflow run. # many narrations fire per workflow run.
[router.narration] [talkie.narration]
temperature = 0.7 temperature = 0.7
top_p = 0.9 top_p = 0.9
max_tokens = 1024 max_tokens = 1024
max_per_run = 100 max_per_run = 100
# Embedder + L3 vector memory. Shared infra: the orchestrator uses these for workflow # Embedder + L3 vector memory. Shared infra: the orchestrator uses these for workflow
# repo-knowledge retrieval, not just the router chat layer — hence top-level, not [router.*]. # repo-knowledge retrieval, not just the Talkie chat layer — hence top-level, not [router.*].
# (Legacy [router.embedder] / [router.l3] are still read as a fallback.) # (Legacy [router.embedder] / [router.l3] are still read as a fallback.)
[embedder] [embedder]
backend = "noop" # or "llamacpp" backend = "noop" # or "llamacpp"