leubeem
76ab765069
Cleanup old unused code and add id sessionId to /api/atc/ptt
2026-05-20 14:13:26 +02:00
itsrubberduck
4a01593612
fix typescript errors and update dependencies
2026-02-17 18:13:04 +01:00
itsrubberduck
cabdc8db6d
fix pm
2026-02-13 08:50:02 +01:00
Remi
d0f3d07200
feat: surface pilot intent from llm decisions
2025-10-16 21:07:21 +02:00
Remi
649cae11bc
feat: integrate llm-backed routing with fallback
2025-10-16 19:55:13 +02:00
Remi
18f25998f5
Allow configuring OpenAI base URL
2025-10-16 11:15:31 +02:00
Remi
79c2cd4870
Simplify decision routing to avoid unnecessary LLM calls
2025-09-25 23:47:05 +02:00
Remi
fd2ac100ce
Align regex auto selection trace with LLM decisions
2025-09-25 23:35:23 +02:00
Remi
60d7d2d89e
Refactor route decision selection
2025-09-25 23:00:31 +02:00
Remi
7f2585c211
Improve flow activation handling and expose active nodes
2025-09-23 23:12:02 +02:00
Remi
fbec5c4830
Add session timeline logging and admin sessions view
2025-09-21 23:08:10 +02:00
Remi
6c9f467b94
Enable flow-aware decision routing
2025-09-21 21:16:33 +02:00
itsrubberduck
137bc84276
stash
2025-09-20 16:36:35 +02:00
itsrubberduck
409f9d5c20
merge
2025-09-20 09:50:20 +02:00
Remi
440aef2f62
Fix remaining German comment
2025-09-20 09:46:34 +02:00
Remi
341a39e7d8
refactor: share llm decision types
2025-09-20 09:27:45 +02:00
Remi
df68719374
refactor: centralize radio speech normalization
2025-09-19 09:20:04 +02:00
Remi
43f2f42238
Add waitlist admin view and log OpenAI decision traces
2025-09-18 23:30:56 +02:00
Remi
74da40cc87
Add readback correction states and improve validation
2025-09-18 19:54:04 +02:00
Remi
8020a4a285
Refine ATC decision routing and ATIS handling
2025-09-18 19:41:32 +02:00
itsrubberduck
8ea2ae28d2
log llm call body
2025-09-18 18:21:01 +02:00
Remi
7a44d67844
Harden runtime config and input validation
2025-09-18 14:03:25 +02:00
itsrubberduck
5ee970e173
even more reduced llm costs but with variables
2025-09-16 13:36:34 +02:00
itsrubberduck
e84feefbe2
feat(openai): Drastically reduce LLM Tokens
...
chore: Optimize the input data sent to the OpenAI API to reduce token usage and improve decision quality. Add quick responses for common pilot utterances like radio checks and emergencies to avoid unnecessary LLM calls.
The key changes are:
- Use a smaller LLM model (`gpt-5-nano`) to reduce token usage
- Optimize the input data by only sending relevant candidate and context information
- Implement quick responses for common pilot utterances like radio checks and emergencies
- Ensure ATC responses are only included when necessary to reduce token usage
2025-09-16 13:18:57 +02:00
itsrubberduck
1e3994f3ce
The system now intelligently handles edge cases while maintaining the core decision tree structure. The LLM will follow the schema when possible but can respond naturally when the pilot says something unexpected, making it much more realistic and flexible for training scenarios
2025-09-16 12:38:15 +02:00
itsrubberduck
f72d5b22b8
new version of pm using decision tree
2025-09-16 12:23:32 +02:00
itsrubberduck
8528b2518b
kann ein bisschen normalized sprechen
2025-09-14 23:20:01 +02:00
itsrubberduck
94ea407402
init
2025-09-14 21:09:47 +02:00
