Changelog

    Back

    Changelog/all //major.minor.patch

    List of all the changes at ⌘ Langbase.

    2024 - DECEMBER

    OpenAI's o1-2024-12-17 model support.

    Web Search with AI Agents guide.

    Cohere's embed-multilingual-v3.0 embedding model support on API.

    Cohere's embed-multilingual-light-v3.0 embedding model support on API.

    Google's text-embedding-004 embedding model support on API.

    Shipped retry mechanism for fetching thread from database

    State of AI Agents 2024 banner on Studio & Docs

    Published 2024 State of AI Agents study

    Llama 3.3 models are now supported by Langbase

    Shipped keyset bar in Pipe IDE

    Shipped memorysets/** → memory/** & added redirects

    Fixed sidebar trigger on mobile view for customer stories

    Pass tools in run API endpoint and SDK

    Fix $ sign being rendered as code snippet in markdown

    Shipped preview URLs for guides, blog, and customer stories

    Blog: Developers Drop GPT-4o Mini in a Flash On Langbase: Google’s Gemini is Faster, Cheaper, Better

    Fixed LLM key error in memory page

    Shipped documentName parameter in memory API endpoints

    2024 - NOVEMBER

    Save LLM tool calls as completion tokens

    Search bar inside memory page

    Shipped tooltips on pricing page

    Shipped API v1 support in Langbase SDK

    Fixed 500 API error if no LLM keys

    Improved profile switcher design

    Shipped docs agent chatbot component

    Improved sidebar design

    Fixed Anthropic model streaming

    Fixed mobile view of changelog all

    Fixed studio favicon

    Shipped syntax highlight in learn content

    Google sign up support

    Fix guides CTA for light mode

    Syntax highlighting for code snippets in Langbase guides

    Improved model selector UI

    Vision support is live on API for different providers

    OpenAI o1 models now support streaming

    gpt-4o-2024-11-20 is available on Langbase

    Improved API rate limits of Pro users to 300 requests per minute

    Shipped sign up ban with disposable email addresses

    Improved pro plan rate limits to 300 requests per minute

    2024 — OCTOBER

    Langbase meet page is now available

    Content infrastructure v2 with central content repository

    Fix model pages, blog, changelog and guides SEO

    xAI models tool call support is now available

    Decrease minimum username length to 4 characters

    Fix object storage endpoint validation

    Fix LLM response parsing when completion and tool call both exists

    Fix token calculation on tool call

    Optimize pipes stats query processing time from 83ms to 6ms

    Tool call support for TogetherAI models that support tool calls

    Tool call support for Anthropic and Google provider models

    Gemini Flash 8B model is now available on Langbase

    Fix username update issues for few users

    Llama 3.2 models are now supported by Langbase.

    API v1 is now available

    Langbase status page

    2024 — SEPTEMBER

    Improved CRUD API performance by 10x using object storage

    Blog: Netlify uses Langbase to bring GeoCities Back to Life

    Update model costs, deprecate Together's deprecated models

    Improved memory deletion

    Run endpoint is live. It replaces the old Generate and Chat endpoints

    Merged Chat and Generate Pipes. Now you can chat and generate in the same pipe

    Support for OpenAI o1 models with handling of its special cases

    Llama 3.1 models are now supported by Langbase.

    Model Pages are now available. One place to get latest details of LLMs

    Improved order of messages in threads

    Langbase Blog

    Fixed docs file icons in memory

    Threads. You can now view all the threads and conversation history of your pipes

    Mistral AI models are now supported by Langbase

    Fix playground timed out for long running requests (>40 secs). New timeout is upto 4 mins

    2024 — AUGUST

    Terms & Conditions and Privacy Policy are now available

    Improve studio sidebar design

    Perplexity models

    Remove unnecessary icon sets and improve icons in studio

    Pricing page get started link now works correctly

    Pipe readme links format properly

    Fork pipe now also fork its functions

    gpt-4o-min as default model in new pipes

    gpt-4o-2024-08-06 model

    JSON mode in Gemini models

    Deprecate qwen72 and add yi-large models

    Explore page ranking algorithm

    Dead links after pipe rename

    Redirect to landing page after pipe delete

    Pricing page is now available

    Preserve period (.) in pipe name

    Promisify sign in page toasts

    Improve studio sidebar watch demo button

    Anchor keysets section in user/org settings

    rag.new and pipe.new redirect fix when logged out

    video=open param to open launch video on landing page

    Internal reserved words

    CMD + S keyboard shortcut on landing page to redirect to sign up

    Pipes page is now available

    Improve onboarding next steps

    Fix error parsing in Google Gemini responses

    Added failed run reason in Usage

    Fix Pipe meta design

    2024 — JULY

    ⌘ Langbase TypeScript SDK is now available

    ⌘ Langbase SDK Examples` for Node.js and Next.js

    ⌘ Langbase SDK generateText() API reference

    ⌘ Langbase SDK streamText() API reference

    Explore page to showcase Open AI pipes

    rag.new to create a new memory on Langbase

    Together Llama 3.1 support

    Fireworks provider support

    Fix JSON mode on gpt-4o-mini

    Together Gemma 2b and 7B support

    GPT-4o mini and Gemma 2 support

    Stream and stream-off code snippets inside Pipe API page

    Stream flag support in API

    Remove waitlist from dashboard

    Update all dashboard dependencies to latest versions

    Auto-redirect to new pipe/memory after creation

    RAG and JSON prompts now switching correctly with every version