apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-07-23 16:51:04 +02:00
Code Issues Projects Releases Packages Wiki Activity

update docs

Browse source
This commit is contained in:
Adil Hafeez 2025-07-11 11:28:03 -07:00
parent af3b38a5b4
commit 0a724ebfd6
No known key found for this signature in database
GPG key ID: 9B18EF7691369645
1 changed files with 15 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

35
README.md
View file

@ -104,10 +104,8 @@ listeners:
timeout: 30s timeout: 30s
llm_providers: llm_providers:
- name: gpt-4o - access_key: $OPENAI_API_KEY
access_key: $OPENAI_API_KEY model: openai/gpt-4o
provider: openai
model: gpt-4o
system_prompt: | system_prompt: |
You are a helpful assistant. You are a helpful assistant.
@ -204,16 +202,12 @@ listeners:
timeout: 30s timeout: 30s
llm_providers: llm_providers:
- name: gpt-4o - access_key: $OPENAI_API_KEY
access_key: $OPENAI_API_KEY model: openai/gpt-4o
provider: openai
model: gpt-4o
default: true default: true
- name: mistral-3b - access_key: $MISTRAL_API_KEY
access_key: $MISTRAL_API_KEY model: mistral/mistral-3b-latest
provider: openai
model: mistral-3b-latest
``` ```
#### Preference-based Routing #### Preference-based Routing
@ -230,17 +224,18 @@ listeners:
timeout: 30s timeout: 30s
llm_providers: llm_providers:
- name: code_generation - model: openai/gpt-4.1
access_key: $OPENAI_API_KEY access_key: $OPENAI_API_KEY
provider_interface: openai default: true
model: gpt-4.1 routing_preferences:
usage: generating new code snippets, functions, or boilerplate based on user prompts or requirements - name: code generation
description: generating new code snippets, functions, or boilerplate based on user prompts or requirements
- name: code_understanding - model: openai/gpt-4o-mini
provider_interface: openai
access_key: $OPENAI_API_KEY access_key: $OPENAI_API_KEY
model: gpt-4o-mini routing_preferences:
usage: understand and explain existing code snippets, functions, or libraries - name: code understanding
description: understand and explain existing code snippets, functions, or libraries
``` ```
Arch uses a lightweight 1.5B autoregressive model to map prompts (and conversation context) to these policies. This approach adapts to intent drift, supports multi-turn conversations, and avoids the brittleness of embedding-based classifiers or manual if/else chains. No retraining is required when adding new models or updating policies — routing is governed entirely by human-readable rules. You can learn more about the design, benchmarks, and methodology behind preference-based routing in our paper: Arch uses a lightweight 1.5B autoregressive model to map prompts (and conversation context) to these policies. This approach adapts to intent drift, supports multi-turn conversations, and avoids the brittleness of embedding-based classifiers or manual if/else chains. No retraining is required when adding new models or updating policies — routing is governed entirely by human-readable rules. You can learn more about the design, benchmarks, and methodology behind preference-based routing in our paper: