A running log of the tools, models, and improvements we've brought to DFIRST.
May
27
Brand DNA Drives the AI Agent + Gemini 3.5 Flash
Your Brand DNA now powers the AI agent end-to-end — go from setup to a fully-branded campaign in a single click. Plus Google's Gemini 3.5 Flash joins the text model lineup with a 1M-token context.
PlatformUX
Brand DNA Powers the AI Agent
Finish your Brand DNA and you're one click away from a fully on-brand campaign. The AI agent now picks up your brand context automatically, generated content matches the language of your company website, and a redesigned main dashboard surfaces a clear "take action" path so your next step is always obvious.
Models
Gemini 3.5 Flash
Google's Gemini 3.5 Flash is now available in the text model picker — a fast, multimodal model with a 1M-token context window, built-in reasoning, function calling, and URL context for pulling in live web pages.
✦
May
20
Brand DNA: Context Toggle & Reset
Brand DNA gets two power-user upgrades — attach it as context to any generation with a single toggle, and reset or regenerate the whole profile when your brand evolves.
PlatformUX
Brand DNA: Toggle as Context, Reset Anytime
Two big upgrades to Brand DNA. First, every generation now offers a single toggle (on by default, surfaced as the first parameter) to attach your Brand DNA as context — so you decide moment-by-moment whether a generation should run with or without your brand profile. Second, you can now delete and fully regenerate your Brand DNA from scratch when your brand evolves, with friendlier social-channel descriptions and a round of reliability fixes throughout.
May
11
Brand DNA, Free Credits & Two New Models
Lock in your brand's voice and style with Brand DNA, earn free credits, get two new top-tier models, and enjoy a much clearer whiteboard — plus mobile and editor polish across the board.
PlatformUX
Brand DNA
Capture your brand's voice, visual style, and personality once — then have it automatically applied across every generation you run in DFIRST. Brand DNA replaces brittle prompt-engineering with a persistent brand profile, so the assets you ship feel on-brand from the first try. Set up during onboarding (and available later in workspace settings) — also accessible to existing customers via a one-time prompt to try it.
PricingUX
Earn Free Credits
Need more credits? You can now earn 2,000 bonus credits (worth $18) by completing a short form, redeemable up to twice a month — no payment required. The new Free Credits entry in the main menu makes it easy to find, and a gentle periodic reminder resurfaces the offer so you don't miss it.
NodesUX
Socket Icons on the Whiteboard
Whiteboard socket connectors now show inline icons for each connection type, so you can see at a glance what plugs into what. Faster wiring, fewer mistakes.
Models
Gemini 3.1 Flash
Google's Gemini 3.1 Flash is now available across text generation tools — a fast, low-latency model well-suited to high-volume drafting and quick iteration.
Models
GPT 5.5
OpenAI's GPT 5.5 is now in the model picker — use it for higher-quality reasoning, longer-form copy, and complex multi-step prompts.
Bug fixes
Bug Fixes
Fixed several bugs across the platform. Most notably: Data Room folder-creation prompts, removing failed nodes from the Feed, output count display during generation, docker scroll after expansion, AI Agent generating overly long prompts, and mobile bugs around the model picker and the credits/settings menu being hidden behind the bottom bar.
✦
Apr
29
Style & Camera Presets in the Docker
Take repeatable control of style, camera, and perspective straight from the docker — without writing technical prompts. Plus a round of fixes across the platform.
UXEditorVideo
Style & Camera Presets
Dial in perspective, camera, lens, lighting, and look-and-feel directly from the docker using semantic presets — front, side, 3/4 or back angles, no technical prompt-engineering required. Presets work consistently across image and video generations, with image-to-image support for even more reliable, repeatable results.
Bug fixes
Bug Fixes
Fixed several bugs across the platform. Most notably: sidebar text tools that always defaulted to a single model, error messages disappearing after a model fallback, Share Node restored for audio nodes, Seedance 1.5 now honoring aspect ratio settings, and the ability to drop .md files onto the whiteboard.
Apr
23
New Models: GPT Image 2, GPT 5.4 & Claude Opus 4.7
Three of the most important model releases of the year landed in DFIRST this week — one new image model and two new flagship text models from the labs setting the pace.
ModelsEditor
GPT Image 2
OpenAI's GPT Image 2 is now available as an image model in DFIRST — with image-reference support and multi-output streaming in the Feed for faster visual exploration.
Models
GPT 5.4 — New Text Model
OpenAI's GPT 5.4 is now a text model option across LLM-powered tools — stronger reasoning, better instruction following, and improved quality on long-form content.
Models
Claude Opus 4.7 — New Text Model
Anthropic's Claude Opus 4.7 — their flagship model — is now live in DFIRST for the most demanding reasoning, agent, and long-context writing tasks.
Bug fixes
Bug Fixes
Fixed a few bugs. Most notably: file upload and the duplicate-and-edit flow in mini-app templates, plus a round of dashboard polish.
✦
Apr
22
Presentations, New Dashboard & Kling 3.0
EditorPlatform
Presentations Generator
Turn any whiteboard node into a polished slide deck. The new presentation tool builds slides from your context, opens them in an in-app editor for manual tweaks, and exports the result as a fully editable PPTX file.
UXPlatform
New Dashboard
A redesigned dashboard greets you with a cleaner layout, live template previews, and faster access to recent projects and ready-made starting points — so you can get from login to first generation in seconds.
UXNodes
Apps Category in the Docker
A new Apps category in the docker puts ready-to-run generators one click away — landing pages, presentations, emails, and assets — each pre-tuned with its own system prompt so you can jump straight from idea to output.
VideoModels
Seedance 2.0
ByteDance's Seedance 2.0 video model is now available, bringing higher-fidelity motion, sharper detail, and stronger prompt adherence for both text-to-video and image-to-video generations.
ModelsEditor
Kling 3.0 Image Omni
Kling 3.0 Image Omni joins the image model lineup — stronger reference handling with multi-image context support for more consistent results across a series.
VideoModels
Kling 3.0 with Motion Control
Kling 3.0 motion control is now available for video generation, giving you finer command over movement and camera dynamics in generated clips.
Bug fixes
Bug Fixes
Fixed a bug where the text sticky note got stuck in a loading placeholder state after being dragged onto the whiteboard.
A full brush tool is now available in the image editor — paint masks, make precise selections, and touch up generations without leaving the canvas.
EditorAudioVideo
Audio Tracks in the Video Editor
The video editor now supports audio tracks — pull sound files from your drive or whiteboard and mix them directly into your videos.
EditorUX
Launch the Editor Without a Whiteboard
You no longer need to open a whiteboard first — launch the image or video editor directly and start working on assets straight away.
Editor
Layer-Preserving Saves
Graphics are now saved with their layer structure intact, so you can come back, tweak one element, and re-export without rebuilding the file from scratch.
EditorUX
Exit Protection with Save Confirmation
Hitting ESC in the editor now prompts to save any pending changes — no more accidentally losing work on a stray keypress.
ModelsUX
Automatic Model Fallback on Errors
When a model fails or is unavailable, DFIRST now suggests a compatible fallback you can switch to with one click — your workflow keeps moving even when a provider hiccups.
ModelsPerformance
Image Streaming for OpenAI & Seedream
OpenAI and Seedream image models now stream results as they generate, so you see progress building up in real time instead of waiting for a final image to appear.
UX
Resizable Docker
The docker panel now has a corner handle you can drag to resize it to fit your screen and your workflow.
UXModels
Favorites Everywhere
Your favorite models now appear front-and-center on the Feed and Whiteboard too — not just in the model selector — so your go-to tools follow you across the app.
VideoModels
Multi-Video Output
Video generations can now produce multiple outputs per run, just like image generations — pick the best take from a batch instead of re-running one at a time.
ResearchUX
Native Scraper Results
Social-media scraper outputs (Instagram, TikTok, LinkedIn, etc.) now render natively inside nodes instead of inside iframes — faster, cleaner, and easier to browse.
Performance
Faster Whiteboards & AI Agent
Deep React optimizations to the whiteboard and AI Agent — noticeably smoother interaction, lower CPU usage, and fewer frame drops on complex projects.
UXPerformance
Rewritten Undo/Redo
Undo and redo now use a proper command-pattern stack — every action is reversible instantly and reliably, even on large projects with heavy history.
PerformanceUX
Virtualized Project List
The side navigation now virtualizes the whiteboard list, so opening a workspace with hundreds of projects is just as fast as opening one with ten.
PerformanceData
Multipart File Uploads
Large file uploads now use multipart transfers instead of base64 encoding — expect dramatically faster uploads for video files, big assets, and Data Room imports.
UXNodes
Faster Docker Start
Opening the docker now puts your cursor straight in the prompt field, and clicking any node instantly opens its configurator — fewer clicks between idea and generation.
Bug fixes
Bug Fixes
Fixed several bugs across the platform. Most notably: Runway Act-Two generation errors, Public Link crashes, asset styling issues, download options restored for Sonar and Gemini Deep Research, audio recordings in the Data Room, and file-rename extension handling.
✦
Mar
23
Shoper Integration & Workspace Rename
IntegrationsPricing
Shoper Integration
Shoper store owners can now install DFIRST directly from the Shoper appstore — accounts are created automatically, plans can be purchased from the Shoper panel, and dedicated pricing is tailored to e-commerce use.
Pricing
Parameter-Based Credit Pricing
Credit costs for generations now reflect the actual parameters you choose — resolution, quality, duration, audio on/off, and more — so you only pay for what you use instead of a flat rate per run.
UXModels
Friendlier Model Selector
Models in the selector now show friendly names, clear descriptions, helpful tags, and expected generation times, making it much easier to pick the right tool for the job.
UXNodes
Unified Edit Tools on the Whiteboard
Base image models and their edit counterparts are now a single tool on the whiteboard. The right mode is picked automatically based on whether you supply a reference — no more switching between Gemini Image and Gemini Image Edit.
UX
Workspace Rename: Projects & Folders
Renamed the workspace structure for clarity — what used to be called a Whiteboard is now a Project, and old Projects are now Folders. Easier to organize, easier to explain.
UXModels
Favorites by Default
The model selector now opens directly to your favorites if you have any saved, so your go-to models are one click away.
Platform
New DFIRST Website
The new public dfirst.ai website is live — featuring an interactive docker so visitors can try the product directly from the landing page before signing up.
Bug fixes
Bug Fixes
Fixed several bugs. Most notably: Kling validation and missing-prompt errors, feed backspace crashes, moving whiteboards between projects, reference sockets disappearing after deselecting a node, and Strategy Hub sections that failed to generate.
Mar
05
Runway Act-Two, Gemini 3.1 Flash & Login Flow
VideoModels
Runway Act-Two Integration
Runway Act-Two is now integrated for video generation, bringing Runway's latest performance-capture model directly into your whiteboard — animate characters with captured human motion.
Models
Gemini 3.1 Flash
Google's Gemini 3.1 Flash is now available for image generation, replacing the deprecated Gemini 3 Pro Preview with a faster, production-ready model.
UX
Cleaner App Login
Visiting app.dfirst.ai without being signed in now takes you straight to the login screen instead of a stale public landing page.
Bug fixes
Bug Fixes
Fixed several bugs. Most notably: lost changes during deploy, YouTube node flipping to transcript on pan, video animations not loading on the whiteboard, drag-and-drop imports in Storage Room, and Deep Research scroll + source link behavior.
Mar
03
New Models: Recraft V4, Gemini 3.1 Pro, Claude Sonnet 4.6 & Seedream 5.0
ModelsEditor
Recraft V4 — New Image Model
Recraft V4 is now available, bringing its latest image generation quality — strong style control, sharp details, and excellent prompt adherence — into your DFIRST workflows.
Models
Gemini 3.1 Pro Preview
Google's latest frontier model, Gemini 3.1 Pro Preview, is now available inside DFIRST for advanced reasoning, long-context, and multimodal tasks.
Models
Claude Sonnet 4.6
Claude Sonnet 4.6 from Anthropic is now live in DFIRST — faster, sharper reasoning and higher-quality writing for content generation and agent workflows.
Models
Seedream 5.0
Added Seedream 5.0, the newest generation of ByteDance's image model, for sharper and more coherent visual output.
Bug fixes
Bug Fixes
Fixed several bugs and added MP3 file support. Most notably: corrected credit cost system, restored PDF downloads from the whiteboard, re-enabled the Edit Image AI tool, fixed Recraft errors on non-1:1 aspect ratios, HTML file imports, and video editor TRIM crashes.
✦
Feb
20
Credit Top-Up Packs
Pricing
Buy Credits Anytime, No Expiry
Running out of credits mid-generation? You can now top up with one-time packs (1,000 / 5,000 / 12,000 credits) that never expire. Plan credits are used first, a combined total is always visible in the menu, and a clear upsell prompt appears the moment you run out — no need to leave your workflow.
Bug fixes
Bug Fixes
Fixed several bugs. Most notably: graphic files not filling the full node, Data Room graphic nodes not showing or loading slowly, music tools broken on Public Link view, unexpected switching from organization to personal workspace, and ElevenLabs audio isolation errors.
Feb
18
Rebuilt Generation Experience, Multi-Node Editing & Deep Research
A rebuilt generation experience, new batch-editing flows, Gemini Deep Research, and Claude 4.6 — a big day across the platform.
PlatformUX
Simplified Grid Generation View
A new generation view that works as a simple grid — great for quick creation — with a one-click toggle to jump into the full whiteboard when you need more control.
UXNodes
Redesigned Tool Configurator
Every tool now uses a unified configurator that works the same way in the simplified view and on the whiteboard. Learn it once, use it everywhere.
UX
Clearer Error Messages
When a generation fails, DFIRST now always explains why — every tool is required to return a human-readable reason, not just a generic "something went wrong."
EditorNodes
Multi-Node Editing
Combine multiple videos and images from your whiteboard into a single edit — select several nodes and open them together in the editor, or pull additional board assets into an active edit without restarting.
ResearchModels
Gemini Deep Research
Gemini Deep Research is now available in DFIRST, letting you run in-depth, multi-source research tasks directly on your whiteboard — with sourced citations and full research trails.
Models
Claude 4.6
Anthropic's Claude 4.6 is now available — improved reasoning, longer context, and better instruction following for agents and content workflows.
Bug fixes
Bug Fixes
Fixed a round of 500 errors on UAT and a bug where graphic files couldn't be opened after download.
Jan
26
Asset Library, New Tools & OmniHuman
DataUX
Asset Library
A central place to browse, reuse, and download every image, video, and sound you have generated — now fully integrated into the whiteboard so you can drop saved assets directly onto any canvas.
VideoModels
Seedance 1.5 Pro
State-of-the-art video generation from ByteDance — Seedance 1.5 Pro delivers higher-quality, more temporally consistent video than the previous generation.
VideoModels
OmniHuman — Portrait to Performance
Added ByteDance's OmniHuman — turn a single portrait plus audio into a realistic talking, singing, or performing human video.
Editor
Remove Background
One-click background removal for any image node — clean cutouts ready to drop onto a new canvas, composite, or product shot.
Editor
Upscale Image
A dedicated image upscaler tool that increases resolution while sharpening details — make smaller generations print- and hero-ready.
VideoEditor
Upscale Video
Boost the resolution and quality of any video node without re-generating from scratch — perfect for getting 1080p or 4K output from faster models.
Editor
Outpainting
Extend any image beyond its original borders — great for reframing portraits, adding breathing room around products, or turning square compositions into widescreen.
Audio
Music Generation
Generate original music tracks on the whiteboard — full songs in any genre from a simple prompt.
AudioVideo
Music for Video
Generate a matching soundtrack for any video node — music that follows the mood and pacing of your footage.
Audio
Voice Generator
Create realistic AI voices in multiple languages and styles — narration, characters, or product voiceovers, all directly on the whiteboard.
AudioVideo
Sound FX for Video
Add AI-generated sound effects to any video node — footsteps, ambient noise, impacts, and more, matched to the footage.
UXNodes
Edit Button on Every Node
Every node on the whiteboard now has an Edit button, giving you one-click access to the right editor (image, video, text, or audio) for fast iteration without hunting through menus.
DataUX
Data Room File Swap
You can now swap the source file inside a Data Room node with a single click, making it easy to rerun the same workflow against different documents without rebuilding your whiteboard.
Pricing
Account Sharing Protection
New protection against multi-user login from different IPs on a single account, plus optimized pricing and billing screens for faster navigation.
PlatformPerformance
Infrastructure Upgrade
A full rebuild of our server infrastructure — faster response times, higher reliability, and the foundation for everything we are shipping next.
Bug fixes
Bug Fixes
Fixed several bugs. Most notably: "Access Denied" error on file delete from storage, Instagram scraper previews, personal space incorrectly showing for organization accounts, HTML downloads, decimal-credit display on regenerate, and Gemini 2.5 image generations getting stuck loading.
✦
Jan
12
Products Module & Gemini 3 Flash
PlatformData
Products Module
A brand-new Products module lets you build your catalog by adding items manually or importing a whole feed from CSV, XLSX, or XML — with product image URL support and a friendly empty-state wizard that guides you from zero to your first product in seconds.
Models
Gemini 3 Flash
Google's Gemini 3 Flash is now a text model option — blazing-fast responses with improved reasoning for everyday prompts and high-volume workflows.
Bug fixes
Bug Fixes
Fixed several bugs. Most notably: Gemini Image aspect ratio handling, Flux Context Pro generation, node group selection and auto-lock, preview node auto-expansion, paid-plan unlocks carrying over, Safari menu overlap on the whiteboard, and AI Agent crashes on UAT.