31 Commits

Author	SHA1	Message	Date
Waleed Latif	7b5168fdd6	feat: added groq as a provider	2025-03-03 14:00:32 -08:00
Waleed Latif	ae226e977c	fix(agent): add dropdown for specific tool selection for blocks that have more than one tool, give that tool to the agent instead of arbitrarily selecting the first tool	2025-03-01 01:04:18 -08:00
Waleed Latif	f97cfddcc1	fix(providers): remove flag to skip proxy for external API calls	2025-02-28 09:16:57 -08:00
Emir Karabeg	07aebc25df	fix(openai): removed parameter mode tool calling	2025-02-28 01:23:12 -08:00
Emir Karabeg	9a05faa170	improvement(tool-calling): adding parameter-only mode	2025-02-27 17:33:17 -08:00
Emir Karabeg	b12652a251	added first request start time	2025-02-27 17:19:51 -08:00
Emir Karabeg	d7285940ac	added openai timer	2025-02-27 17:15:34 -08:00
Waleed Latif	9fa0551179	feat(providers/cerebras): use dedicated cerebras sdk	2025-02-27 16:06:20 -08:00
Waleed Latif	16f5065678	feat(custom-functions): add support for custom functions, tested with & without code provided	2025-02-27 15:03:30 -08:00
Waleed Latif	5c11e9da16	feat(providers): modified all providers to use SDK rather than creating an HTTP request	2025-02-27 13:39:33 -08:00
Waleed Latif	a08efa3d91	feat(providers): added openai-sdk in place of custom business logic	2025-02-27 13:29:52 -08:00
Waleed Latif	fa67494fe7	feat(providers): removed providers from tools directory, added cerebras sdk	2025-02-27 13:29:52 -08:00
Waleed Latif	ee1932c856	feat(models): added claude-3-7-sonnet	2025-02-24 13:35:53 -08:00
Waleed Latif	f6a639302e	fix(starter): add temp logging to debug executor inability to resolve envvars	2025-02-19 13:49:36 -08:00
Waleed Latif	bf6e91fcbb	Remove extra logs from OpenAI provider	2025-02-14 20:16:50 -08:00
Waleed Latif	d99bd1e0a9	Add more thorough check for JSON in providers/	2025-02-13 18:14:58 -08:00
Waleed Latif	f74aa815f3	Cleanup extra, unnecessary logs	2025-02-13 02:28:56 -08:00
Waleed Latif	af323a71ab	Modify executor to treat evaluator as an 'output' block instead of a 'router' block, looping logic still needs to be updated but forward path works	2025-02-13 00:59:25 -08:00
Waleed Latif	2818cd91a9	Added code to trim gemini response of any markdown/whitespace	2025-02-12 13:37:51 -08:00
Waleed Latif	660b8887b0	Updated gemini model to 2.0 flash	2025-02-07 00:19:36 -08:00
Waleed Latif	a10a724a51	Added support for response format in agent block. Works in all but deepseek (will fix) and gemini (does not support for this model). If responseformat is not specified, we get the response back wrapped in 'content' like before	2025-02-06 19:59:44 -08:00
Waleed Latif	4df1c8268c	Removed duplicate logic to map model name onto provider	2025-02-05 18:27:11 -08:00
Waleed Latif	c4559429a3	Fixed typo in tavily tool, acknowledge known error in deepseek-v3 function calling with infinite loops, consolidate providers in registry under providers directory, fixed error handing from 3P APIs	2025-02-05 11:54:19 -08:00
Waleed Latif	9493055c60	Added prettier, added husky pre-commit hook to ensure formatting	2025-02-05 03:10:23 -08:00
Waleed Latif	d3abd849b3	Changed o1-mini to o3-mini, supports calls without function calling now	2025-02-04 17:17:51 -08:00
Waleed Latif	204469d387	Added xai function calling support	2025-02-04 17:06:39 -08:00
Waleed Latif	a823d6e2ea	Added deepseek-v3 function calling	2025-02-04 17:01:00 -08:00
Waleed Latif	a154e8f37e	Added anthropic tool calling functionality	2025-02-04 16:50:34 -08:00
Waleed Latif	5ae95bbed9	Added gemini tool calling support	2025-02-04 15:43:14 -08:00
Waleed Latif	832a35354b	Generalized provider implementation so adding other providers will be easier	2025-02-04 15:31:52 -08:00
Waleed Latif	98c39afe43	Created new providers/ directory for model providers, modified tool names to have '_' instead of '.', added support for function/tool calling by agent block. Works for OpenAI w firecrawl tool, needs to be implemented still for other providers	2025-02-04 13:07:03 -08:00