hermes-agent/website/docs/user-guide/features
Teknium 543d305bbb
feat(moa): add reference_max_tokens to cap advisor output and cut turn latency (#56756)
MoA per-turn latency is dominated by advisor GENERATION: turn wall time
correlates ~0.88 with output tokens and ~-0.03 with input tokens (measured over
52 turns). Each turn waits for the slowest advisor to finish writing, and
advisors were uncapped — writing multi-thousand-token essays the aggregator
only needs the gist of.

Add an opt-in per-preset reference_max_tokens knob (mirrors reference_temperature)
that caps ADVISOR output only; the acting aggregator is never capped. Default
None = uncapped, so existing presets are byte-for-byte unchanged (no regression).
Wired through both MoA execution paths (MoAChatCompletions.create and
aggregate_moa_context).

E2E: same task, closed preset uncapped vs reference_max_tokens=600 -> 59s to 33s
(~44% faster), final answer identical/correct.

- hermes_cli/moa_config.py: _coerce_int_or_none helper + reference_max_tokens
  in _normalize_preset/_default_preset/flattened view
- agent/moa_loop.py: read preset.reference_max_tokens, pass to reference fan-out
- agent/conversation_loop.py: pass reference_max_tokens on the per-turn path
- tests + docs
2026-07-02 00:16:35 -07:00
..
_category_.json
…
acp.md
…
api-server.md
…
batch-processing.md
…
browser.md
…
built-in-plugins.md
…
code-execution.md fix: drain thread no longer crashes on fd-less stdout streams (#34789) 2026-05-29 12:16:57 -07:00
codex-app-server-runtime.md
…
computer-use.md
…
context-files.md
…
context-references.md
…
credential-pools.md
…
cron.md docs(cron): scope in_channel to channels; document DM continuation knob 2026-07-01 03:16:13 -07:00
curator.md
…
delegation.md docs: reconcile docs with code across last 3 releases (#54254) 2026-06-28 12:47:50 -07:00
deliverable-mode.md
…
extending-the-dashboard.md
…
fallback-providers.md
…
goals.md
…
honcho.md
…
hooks.md feat(agent): add pre_verify hook and verify-on-stop coding guidance 2026-06-30 00:59:29 -05:00
image-generation.md
…
kanban-tutorial.md
…
kanban-worker-lanes.md
…
kanban.md
…
lsp.md feat(lsp): add PowerShellEditorServices language server (#55930) 2026-06-30 16:22:18 -07:00
mcp.md docs: reconcile docs with code across last 3 releases (#54254) 2026-06-28 12:47:50 -07:00
memory-providers.md feat(memory): add Supermemory setup connection summary 2026-06-27 15:07:34 +05:30
memory.md
…
mixture-of-agents.md feat(moa): add reference_max_tokens to cap advisor output and cut turn latency (#56756) 2026-07-02 00:16:35 -07:00
overview.md
…
personality.md
…
pets.md docs: reconcile docs with code across last 3 releases (#54254) 2026-06-28 12:47:50 -07:00
plugins.md
…
provider-routing.md
…
skills.md
…
skins.md
…
spotify.md
…
subscription-proxy.md
…
tool-gateway.md
…
tool-search.md docs: reconcile docs with code across last 3 releases (#54254) 2026-06-28 12:47:50 -07:00
tools.md docs: reconcile docs with code across last 3 releases (#54254) 2026-06-28 12:47:50 -07:00
tts.md
…
vision.md
…
voice-mode.md
…
web-dashboard.md docs: reconcile docs with code across last 3 releases (#54254) 2026-06-28 12:47:50 -07:00
web-search.md docs: fix incorrect web search instructions 2026-06-28 02:54:27 -07:00
x-search.md
…