What's new?
- Initial release.
What's new?
- Enhanced response generation
- Prompt optimizers added - [code, shell_command]
- Prompt optimizers added - Console
- Clear console - Console
What's new?
- Busy bar index bug fixed - (Console)
- Other minor fixed.
What's new?
- Minor fixes
What's new?
- Multiple variables renamed.
- First release under
python-tgpt
What's new?
generate
is the default command. Thanks to @sameedzahoor- Control response framing
--quiet
generate
based prompt optimizaion ---code
&--shell
What's new?
- Chat conversationally - (Experimental)
- Maintain chat history .txt
- Load chat history from file
- Chain request through proxies
What's new?
- Reading piped input as prompt - (Console). Thanks to @sameedzahoor
- Reset conversation - (Console)
- View conversation history - (Console)
- Other minor fixes
What's new?
- Chatting conversationally - Stable
What's new?
- Chatting conversationally - Default Mode
What's new?
- Bug fixed - file not found error
What's new?
- Check version -
--version
- Minor fixes.
What's new?
- Invalid response - fixed
What's new?
- Incomplete response - fixed
What's new?
- Multiple LLM providers
What's new?
- Fakeopen rendering issue fixed - #7
What's new?
What's new?
- Tabulated help info -
h
- Control response vertical overflow behaviour.
What's new?
- WebChatGPT added as provider
- Awesome-prompts manipulation commands = CRUD
- Other minor fixes.
What's new?
- New provider : Bard by Google.
- Check whole last response
- Other minor fixes.
What's new?
- Bug fixed - reset conversation - bard
- Bug fixed - low
httpx
logging level. - bard
What's new?
- Busy bar disabled when
--quiet
issued in generate mode. #12 Thanks to @johnd0e interactive
takes action on$ pytgpt
otherwise one has to explictly declare the action. #11
What's new?
- Auto-quiet on output redirection. Thanks to @johnd0e
- Dropped support for sourcing prompt from
stdin stream
in Windows. #12 - Colorized command prompt. interactive
What's new?
- Improved command prompt - icon & color blending
- Bug fixed - multiline prompt in
interactive
what's new?
- Improved introductory prompt
You're a Large Language Model for chatting with people. Assume role of the LLM and give your response.
- Combine both of piped and explicitly issued prompt #13
- Support piping input in Windows. #13
- Placeholder for piped input
{{stream}}
and copied text{{copied}}
.
What's new?
41 New models. Thanks to gpt4free.
- AiChatOnline
- Aura
- Bard
- Bing
- ChatBase
- ChatForAi
- Chatgpt4Online
- ChatgptAi
- ChatgptDemo
- ChatgptNext
- Chatxyz
- DeepInfra
- FakeGpt
- FreeChatgpt
- GPTalk
- GeekGpt
- GeminiProChat
- Gpt6
- GptChatly
- GptForLove
- GptGo
- GptTalkRu
- Hashnode
- HuggingChat
- Koala
- Liaobots
- Llama2
- MyShell
- OnlineGpt
- OpenaiChat
- PerplexityAi
- Phind
- Pi
- Poe
- Raycast
- TalkAi
- Theb
- ThebApi
- You
- Yqcloud
- Aura is the default provider
- Other minor fixes.
What's new?
- Added utility commands
utils
gpt4free
update
etc - Support g4f-based non-working providers.
- Drop restriction to specific
g4f
dependency version (#14 (comment)) - Pass auth value to g4f-based providers.
-k
- Support proxying in g4f providers
What's new?
- gpt4free gui interface -
web
FakeGpt
- default provider.- Other minor updates.
What's new?
- Auto-detect placeholders (#17)
Aura
- default provider
What's new?
- Binaries for all system.
What's new?
- New provider - Phind . Resolves #18
- New util command
latest
. phind
is the default provider - console- Fixed :
opengpt
- (#19)
What's new?
- Binaries dependencies fixed.
What's new?
- Phind fixed. #21
- Bard fixed. #23 #15
- Full and minimal executables.
What's new?
- Phind fixed. ##21 (comment)
What's new?
- New provider - Llama2
- New provider - Blackboxai
- Disable g4f version check.
- Partial or full installation options.
What's new?
- Rawdog : generate and execute python code in your system, driven by your prompts.
What's new?
- RawDog:
- Give consent to script execution
- Execute script internally or externally
- Choose python interpreter name
What's new?
- Minor bug fix. RawDog : generate
What's new?
- Execute python code in responses (interactive)-
exec
- Execute python codes using system installed python interpreter - default
- Other minor fixes.
What's new?
- New model : GPT4ALL - Support offline LLM.
What's new?
- Revamped provider
webchatgpt
. - Dynamic provider
g4fauto
. #29 - Test and save working g4f providers . #29
pytgpt gpt4free test -y
- Order providers in ascending. #31
What's new?
g4fauto
fixed.
What's new?
- Execute scripts native interpreter - rawdog
- Typos fixed.
- Other minor fixes.
What's new?
webchatgpt
supports intro prompt.- Fix: placeholders consistency -
{{stream}}|{{copied}}
- Other minor fixes.
What's new?
- added: Provider poe. Supports multiple models. Thanks to snowby666/poe-api-wrapper.
What's new?
- added: Load variables from
.env
file. - patch: Exclude selenium dependents providers from test. #33
- patch: Show more gpt4free models.
- added: Test logging flag.
What's new?
- fix: llama2. #34
What's new?
- added: Support for image generation.
What's new?
- fix: Gemini (Bard) provider.
- added: awesome prompts info.
What's new?
- patch:
FreeGpt
default g4f-based provider. - added:
g4f
made required dependency. - fix:
openai
response repetition. #39 - Other minor fixes.
What's New?
- added: Provider Groq
What's New?
- fix: Provider
Opengpt
.
What's New?
- added: Provider
perplexity
. Thanks to HelpingAI/Helpingai_T2
What's new?
- added : FastAPI.
- New extra -
api
What's new?
- added : FastAPI - Image generation
- added : FastAPI - Providers endpoint.
- patch : FastAPI - existing endpoints
What's new?
- patch : FastAPI -
/images
amount > 10 causing http status_code 500. - added : Hosted api health-status monitor.
What's new?
- added : API static contents -
clear
command.$ pytgpt api clear images
- patch : Streaming redirected outputs.
pytgpt generate --raw "Write a short story" > shortstory.txt
#43
- other minor updates
What's new?
- feat : Dynamic provider -
auto
: Working provider overally. - Other minor feats.
What's new?
- feat: New text provider - YepChat
- feat: New image provider Prodia
- feat: Speech synthesise responses.
--talk-to-me
shortform-ttm
- feat: Speech synthesise - FastAPI -
/audio
What's new?
- feat: Minor mods
- addon: Telegram-bot - pytgpt-bot
What's new?
- feat: Override chat intro.
- Other minor fixes.
What's new?
- fix: Failure to include intro in chat history - from
v0.6.6
What's new?
- fix: Improper incomplete chat history generation.
- feat: Load intro from chat history file. (first line)
- feat: Set new
intro
on the fly. (console) - feat:
en-US-Wavenet-C
- default voice for speech synthesis. - fix: Perplexity AI raising
json.decoder.JSONDecodeError
What's new?
Note
This is a remarkable milestone. We have to do it in style.
- feat: Asynchronous implementation to all providers except a few.
- feat: Asynchronus implementation to all
FastAPI
endpoints.
What's new?
- fix: Failure to support g4f providers -
FastApi
- feat: Rename
/audio
endpoints to/voice
-FastAPI
What's new?
- feat: Extra installation requirement for Termux -
pip install python-tgpt[termux]
- fix: Limit g4f requirement for binaries to v0.2.6.1
What's new?
- feat: Termux extra installation options :
termux-cli
,termux-api
andtermux-all
What's new?
- fix : Image generation failure.
What's new?
- feat: New provider - Novita
What's new?
- feat: Added prompt completions in interactive mode
cli
What's new?
- fix: Remove providers: yepchat, opengpt, leo etc