Commit History

  1. auto-fallback: try InferencePort, if 429 use KeylessAI

    ausername-12345
    26ac8e0
  2. switch to InferencePort direct API

    ausername-12345
    3afc934
  3. back to KeylessAI (completions.me 401)

    ausername-12345
    d6ecf4e
  4. read API key from env var with fallback

    ausername-12345
    cfe2661
  5. switch to completions.me with gpt-5.2

    ausername-12345
    b854661
  6. compact history to 20 messages: older ones get summarized into context

    ausername-12345
    ca7c777
  7. log all requests with full details

    ausername-12345
    803f11d
  8. switch to KeylessAI (no key, no limits, no signup)

    ausername-12345
    eb7395f
  9. add web_type tool for filling forms and typing in textboxes

    ausername-12345
    65e6ab8
  10. add web_search, web_click tools for interactive browsing

    ausername-12345
    75dd203
  11. add web browsing: web_fetch tool + tool call loop

    ausername-12345
    71a2b74
  12. restore original request parsing for shortcuts + history

    ausername-12345
    f171792
  13. fix shortcuts empty history, add debug logging

    ausername-12345
    1c3e0d8
  14. add GET /history, handle messages array in shortcuts

    ausername-12345
    dce2588
  15. separate POST /reset endpoint for clearing history

    ausername-12345
    346f0d8
  16. shortcuts maintains conversation history, reset with {"reset":true}

    ausername-12345
    987e29f

  1. rename relay_history to history

    ausername-12345
    f2448b2
  2. restore relay_history as string

    ausername-12345
    8e9bc3f
  3. remove relay_history from response

    ausername-12345
    af6863a
  4. relay_history as string

    ausername-12345
    5a42964
  5. relay_history excludes system prompt; API still sends it

    ausername-12345
    144b36d
  6. shortcuts: send only messages, no system prompt

    ausername-12345
    b69140b
  7. shortcuts: send all messages without trimming

    ausername-12345
    8e74d92
  8. add relay_history to response

    ausername-12345
    fc7e0fe
  9. add request logging on failure (method, url, headers, body)

    ausername-12345
    39aab2b
  10. restore default system prompt for /shortcuts

    ausername-12345
    5856d7c
  11. make /shortcuts accept simpler formats (text/message)

    ausername-12345
    bdb7f72
  12. add POST /shortcuts endpoint (alias for /relay)

    ausername-12345
    872ca3e

  1. use lightning model (Lightning-Text v2)

    ausername-12345
    ed6fce5

  1. fix: use llama-3.1-8b-instant model (llama3-8b-8192 decommissioned)

    ausername-12345
    7a0bf02
  2. sync with sebebeb-relay.hf.space: fix API URL to /v1/chat/completions, add debug endpoint, restore OpenAI payload format

    ausername-12345
    0a988dd