AnythingLLM v.1.2.2 | The custom agent skill release!

🥳 Big Changes

• Custom Agent Skills: Create fully custom agent skills to extend the capabilities of your @agent invocations. If it can be done in NodeJS - you can do it here.

• AnythingLLM Browser Extension: Send and embed information from the web directly to AnythingLLM or embed directly into your workspaces for later! This can allow you to even scrape data from password protected, VPN protected, or otherwise non-public webpages

• You can now use @agent to run skills via the developer API. (nice)

What's Changed

• Update common.js by @hejp001 in #2148
• Refactor api endpoint chat handler to its own function by @timothycarambat in #2157
• Api session id support by @timothycarambat in #2158
• Enabled use of @agent (and skills) via dev API calls by @timothycarambat in #2161
• Move embed to submodule by @timothycarambat in #2163
• Update common.js by @hejp001 in #2165
• Add RAG agent plugin to API agent by @timothycarambat in #2171
• feat: Added Elestio as one-click deploy option by @kaiwalyakoparkar in #2168
• Update common.js by @hejp001 in #2179
• Support gpt-4o for Azure deployments by @timothycarambat in #2182
• Mobile render fix by @jeffscottward in #2176
• AnythingLLM Chrome Extension by @timothycarambat in #2066
• handle onblur for emebed domains by @timothycarambat in #2199
• Add new Voyage AI embedding models by @shatfield4 in #2193
• Bug/make swagger json output openapi 3 compliant by @MrSimonC in #2219
• Feature/use escape key to close documents modal by @timothycarambat in #2222
• Feature/add searchapi web browsing by @timothycarambat in #2224
• fixed the typo in LLMs by @amrrs in #2225
• Ollama sequential embedding by @timothycarambat in #2230
• remove Jazzicons & Add default pfps by @timothycarambat in #2233
• update docs showing no need for manual port forwarding of Server in G… by @MrSimonC in #2247
• match user prompts exactly not partially by @MrSimonC in #2245
• Milvus bug fix by @shatfield4 in #2183
• Fix UI for slash cmd presets by @shatfield4 in #2260
• Add support for custom agent skills via plugins by @timothycarambat in #2202
• Add Gemini exp models by @timothycarambat in #2268
• Update OpenAI models and prices #2261 by @timothycarambat in #2269
• UI Cleanup by @timothycarambat in #2270

New Contributors

• @hejp001 made their first contribution in #2148
• @kaiwalyakoparkar made their first contribution in #2168
• @jeffscottward made their first contribution in #2176
• @amrrs made their first contribution in #2225

Full Changelog: v1.2.1...v1.2.2

this version correlates with desktop v.1.6.5

AnythingLLM v1.2.1

What's Changed

• Slash commands edit bug fix by @shatfield4 in #2073
• Populate .env file using variable inside userdata script by @saravanan30erd in #2075
• add exp model to v1Beta by @timothycarambat in #2082
• #2084 Support Go filetype by @leaanthony in #2085
• Update the display to show 95% of the width on small screens, else 635px by @MrSimonC in #2089
• Fix depth handling in bulk link scraper by @shatfield4 in #2096
• Finetune ui improvements by @shatfield4 in #2053
• Improve search of document picker by @shatfield4 in #2097
• Auto URL detection for localai by @shatfield4 in #2083
• Add prompt window limits for gpt-4o-* models by @mPyKen in #2104
• fix(Dockerfile): remove hardcoded exposed port by @kabaluyot in #2106
• Encryption in JWT for single-user password mode by @timothycarambat in #2111
• Fix TypeError by replacing this.openai.createChatCompletion with the correct function call by @shaharhi in #2117
• Agent Context window + context window refactor. by @timothycarambat in #2126
• Fix markdown styling for rendered images by @shatfield4 in #2125
• Fix tooltip showing on both sides of file picker by @shatfield4 in #2124
• Omit invalid response.text values and prompts by @timothycarambat in #2127
• Fix multi-user setup bug for invalid username/password by @shatfield4 in #2130
• Allow readable username passed in script widget by @timothycarambat in #2131
• Support suggestions of messages for empty chat in widget by @timothycarambat in #2133
• Added Hebrew language by @timothycarambat in #2143

New Contributors

• @saravanan30erd made their first contribution in #2075
• @leaanthony made their first contribution in #2085
• @mPyKen made their first contribution in #2104
• @kabaluyot made their first contribution in #2106
• @shaharhi made their first contribution in #2117

Full Changelog: v1.2.0...v1.2.1

Desktop Version Pin: v1.6.4

AnythingLLM v1.2.0

New features

• Full PiperTTS for in-browser text-to-speech with over 100+ voices (s/o to @rhasspy & @diffusion-studio for MIT work & base-fork of NPM package)
• Full Portuguese by @im-ramon
• Full German Translation
• Multiple /slash command expansion in prompts by @shatfield4
• MathJax/KaTeX/LaTeX support in responses

Fixes & others

• Fix typo by @martincenturion in #2041
• 2034 automate port forwarding for GitHub codespaces using gh cli by @MrSimonC in #2035
• Youtube loader whitespace fix by @shatfield4 in #2051
• 2049 remove break that prevents fetching files from gitlab repo by @mehmetunlu in #2050
• Hide upload element for default role users by @timothycarambat in #2061
• New user account validations by @shatfield4 in #2037
• Developer API access for embed chat widgets by @shatfield4 in #1999
• Speech to text hotkey by @Rrojaski in #1771

New Contributors

• @martincenturion made their first contribution in #2041
• @MrSimonC made their first contribution in #2035
• @im-ramon made their first contribution in #2054
• @mehmetunlu made their first contribution in #2050
• @Rrojaski made their first contribution in #1771

Desktop version pin: v1.6.1

Full Changelog: v1.1.1...v1.2.0

AnythingLLM v1.1.1

What's Changed

• Gemini Pro 1.5, API support for 2M context and new experimental model by @RahSwe in #2031
• Added Italian translation by @rglauco in #2025
• handle OpenRouter exceptions on streaming by @timothycarambat in #2033
• Ollama performance mode option by @shatfield4 in #2014

New Contributors

• @RahSwe made their first contribution in #2031
• @rglauco made their first contribution in #2025

Full Changelog: v1.1.0...v1.1.1

AnythingLLM v1.1.0 w/multimodality support

What's Changed

Notable

• Add multimodality support by @timothycarambat in #2001
• Add drag-and-drop to chat window for auto upload & embed of documents by @timothycarambat in #1995

---

• Updated API for workspace to return active threads by @timothycarambat in #1978
• Fix GroqAI picker UI bug by @timothycarambat in #1993
• Enable editing of OpenRouter stream timeout for slower connections by @timothycarambat in #1994
• Drag and drop file colors update & margin styles by @shatfield4 in #2000
• Added more char exceptions to mapping extension by @timothycarambat in #2010

Full Changelog: v1.0.0...v1.1.0

AnythingLLM v1.0.0 🎉

Hey everyone, it's high time we start a formal release cycle for AnythingLLM. As such, we will now be tagging releases as we have some both large and smaller features that will require version tracking.

---

What's Changed

• Implement Chroma Support by @timothycarambat in #1
• Lancedb support by @timothycarambat in #6
• Update package.jsons to reflect new project name by @timothycarambat in #11
• 12 auth implementation by @timothycarambat in #13
• Change default server port since 5000 can be taken by default. by @timothycarambat in #15
• fix input color in non-dark mode screens by @timothycarambat in #17
• Fix pypandoc issue by @hipnologo in #23
• Fix small react and JS errors by @timothyasp in #40
• Fix link to Vector Caching in README.md by @m14t in #38
• update readme with build instructions by @timothyasp in #41
• Docker support by @frasergr in #34
• refactor: convert insert loop to 1 insert stmt by @jwaltz in #19
• Add support for CPU arch in Docker installation by @preeteshjain in #48
• Easy AWS cloud deployment for private instances via CloudFormation template by @timothycarambat in #52
• Adds ability to import sitemaps to include a website by @skidvis in #51
• add codeblock support for prompt replies and historical messages by @timothycarambat in #55
• Blocks images in sitemaps from being parsed. by @skidvis in #56
• Implement Workspace-specific setting configs + other technical features by @timothycarambat in #58
• Add support for all CPU architectures in docker compose by @preeteshjain in #60
• ignore storage folder changes for nodemon by @timothycarambat in #66
• Update HOTDIR.md by @AntonioCiolino in #70
• Upload and process documents via UI + document processor in docker image by @timothycarambat in #65
• Enable web scraping based on a url and a simple filter. by @AntonioCiolino in #73
• Add mobile stylings to UI by @timothycarambat in #75
• Dockerfile cleanup and enforce Unix line endings by @frasergr in #81
• updated instructions in docker readme by @frasergr in #79
• Updated Link scraper to avoid NoneType error. by @AntonioCiolino in #90
• Added mbox support by @mplawner in #106
• Enable debug mode of ENV at runtime by @timothycarambat in #111
• Add chat/converstaion mode as the default chat mode for all Vector Databases by @timothycarambat in #112
• add ability to purge document from custom documents as well as cleanup its associated cache file by @timothycarambat in #113
• dump error message to frontend instead of generic error by @timothycarambat in #114
• History per workspace by @AntonioCiolino in #126
• be able to parse relative and FQDN links from root reliabily by @timothycarambat in #138
• Twitter Feature by @AntonioCiolino in #134
• Adding GCP Deployment Template by @ajosegun in #142
• Enabled importing and exporting of entire AnythingLLM instance data by @timothycarambat in #146
• Adding Digital Ocean Deployment using Terraform by @ajosegun in #148
• [Fork] Additions on franzbischoff resolution on #122 by @timothycarambat in #152
• [Fork] Batch embed by jwaltz by @timothycarambat in #153
• LanceDB upgrade and improved functionality. by @timothycarambat in #155
• Enable the system owner to be able to update the system wide password and secret by @timothycarambat in #156
• [FEATURE] Enable the ability to have multi user instances by @timothycarambat in #158
• improve citations to show all text chunks referred and expand the citation to view full referenced text by @timothycarambat in #161
• Move OpenAI api calls into its own interface/Class by @timothycarambat in #162
• better vector db selection UI on settings by @timothycarambat in #175
• Split large PDFS into subfolder in documents by @timothycarambat in #176
• Multiple LLM Support framework + AzureOpenAI Support by @timothycarambat in #180
• Actually enable x-scrolling when citations overflow. by @efocht in #178
• Add support for Weaviate VectorDB by @timothycarambat in #181
• Solving permissions issue with collector/outputs and server/storage by @efocht in #182
• Update requirements.txt by @melroy89 in #185
• Logo customization for single-user & multi-user modes by @shatfield4 in #186
• Enable optional Telemetry integration via posthog by @timothycarambat in #190
• Add Qdrant support for embedding, chat, and conversation by @timothycarambat in #192
• Added support for host.docker.internal (ie: to use local chromadb ser… by @jflaflamme in #195
• Custom default messages implementation for single and multi-user modes by @shatfield4 in #193
• improve UX and useability for Settings and prefernces for mobile and desktop by @timothycarambat in #200
• Implement dumpENV if in prod environment by @shatfield4 in #207
• Autoscroll to end of chat fix by @shatfield4 in #201
• Enable batch deletion of Pinecone Ids by max limit of 1000 by @timothycarambat in #210
• Make openAI Azure embedding requests run concurrently to avoid input limits per call by @timothycarambat in #211
• Success fail messages for upload document by @shatfield4 in #208
• Implement toast for upload event by @shatfield4 in #214
• Full developer api by @timothycarambat in #221
• fix settings overlay hiding underneath prompt by @timothycarambat in #222
• Implemented toast messages for all system settings by @shatfield4 in #219
• Hide delete workspace button on non-admin users if setting is disabled by @shatfield4 in #227
• prevent exports path traversal by @timothycarambat in #233
• patch SQL injection opportunities [LOW RISK] by @timothycarambat in #234
• Adding messages to embedding updates by @shatfield4 in #229
• Improve UI for login modal by @shatfield4 in #235
• Franzbischoff document improvements by @timothycarambat in #241
• Display better error messages from document processor by @timothycarambat in #243
• Renamed all indicies to vectors to avoid confusion of vocab by @shatfield4 in #246
• changing the build behavior in the aws cloudformation template from [… by @santini-cyber in https://github.com//pull/247
• Replace custom sqlite dbms with prisma by @shatfield4 in #239
• Prevent updates of specific keys via API by @t...