Eddie AI Overhauled – Pay-As-You-Go Pricing, Editing from Claude and ChatGPT, and an iPhone App

Software News

by Johnnie BehiriAugust 28^th, 2026

Subscribe to News - Software

Eddie AI has rolled out its broadest set of changes since the v3 release, and the most consequential one is not a feature at all: subscriptions are gone. The assistant editor now runs on pay-as-you-go credits, exposes its entire toolset to Claude and ChatGPT through an MCP server, and fits in your pocket with a native iPhone app. Add generative music, voiceover, and b-roll to the list, and the pitch shifts from a product you open to a service you reach from wherever you already work.

The update touches pricing, access, and output in one sweep. According to Eddie AI, the platform is now often around 50% cheaper under the new credit model, which starts every account with 1,000 free credits and sells packs from $10 for 1,000 credits with no monthly fee attached. Around a hundred MCP tools cover transcripts, soundbites, assembly, b-roll, captions, music, voiceover, multicam angles, and NLE export, so an edit can be directed entirely from a chat window. The free iPhone app handles prompting and review on the go, cloud workspaces bring Slack, Microsoft Teams, and enterprise SSO into the mix, and the company will show all of it at IBC 2026 in Amsterdam. If you are new to the tool, our NAB 2026 hands-on with the Night Shift workflow is a good primer on what Eddie AI actually does with raw footage.

Subscriptions are out, credits are in

The pricing change is the structural one. Eddie AI previously ran on subscription tiers alongside a per-credit option, a model we last looked at in our conversation with CEO Shamir Allibhai at Cine Gear. That structure is gone. The pricing page now lists credit packs only: $10 for 1,000 credits, $45 for 5,000, $120 for 15,000, and $350 for 50,000, with custom enterprise pricing above that. Credits never expire, every feature is available regardless of pack size, and auto-recharge is optional and switched off by default, a small but welcome detail in a market where metered AI services tend to assume the opposite.

Credits meter everything the system does: importing footage runs about 500 credits per hour of source material, billed in 15-minute blocks, while edit runs are priced on the work involved, and larger jobs show a cost estimate before they execute. The company’s claim that this works out to often 50% less than before is hard to verify directly, since the old and new credit systems are not the same unit. What can be said is that the entry barrier has dropped to zero: every new account starts with 1,000 free credits, enough to take a small project from import to export before spending anything.

Cutting footage from Claude and ChatGPT

The second pillar is distribution. Eddie AI now runs as an MCP server, meaning the Model Context Protocol clients many editors already have open, Claude and ChatGPT among them, can drive an edit directly. The company says around a hundred tools are exposed this way, spanning transcripts, soundbite selection, assembly, b-roll placement, captions, music, voiceover, multicam angle switching, and export to the NLE. In practice, you describe the cut in the chat interface you already use, and Eddie AI executes it against the footage sitting in your cloud project.

Inside ChatGPT, the result comes back as a playable card built from real frames of your footage rather than a link out to another app. The workflow logic is worth pausing on: the language model never touches the media itself. It plans and directs, while Eddie AI holds the footage, the timecode, and the timeline, which is the part a chat assistant cannot do on its own. For editors who finish in Premiere Pro, DaVinci Resolve, or Final Cut Pro, the export path remains the same one we covered when Dirty Multicam support landed: Eddie AI hands over a timeline, and the finishing pass stays human.

An iPhone app for directing the edit, not making it

Eddie AI’s native iPhone app is now on the App Store as a free download requiring iOS 15.1 or later. The scope is deliberately narrow. You prompt an edit, watch the finished cut in a gapless player, save the MP4 to Photos, or share a link that anyone can open in a browser without an account. Footage imports from the camera roll, Google Drive, or Frame.io, and a push notification tells you when a job is done, so the intended pattern is to start the work and put the phone away.

The company is explicit that this is not a mobile editing suite. The heavy media stays in the cloud, and the app covers review and direction: the producer checking a cut from set, the editor approving a revision between locations. Projects are the same cloud projects used on web and desktop, so a change prompted from a phone shows up on the workstation timeline.

Cloud projects, workspaces, and frame-accurate comments

Underneath the new endpoints sits a reworked collaboration layer. Projects live in the cloud and sync live across web, desktop, and phone. Workspaces give a team shared membership, shared projects, and a shared credit pool, and share links now carry real access levels instead of a single view-only mode. Comments land on the exact frame and push to the editor instantly, extending the direction Eddie AI started with its timestamped Feedback Mode v2 earlier this year.

For larger operations, Slack and Microsoft Teams integrations put project activity into the channel a team already watches, and enterprise SSO is now available. Storage connections cover Google Drive, Dropbox, Frame.io, iconik, and LucidLink, which reads like a deliberate courtship of broadcast and agency post departments, not solo creators.

Generative music, voiceover, and b-roll

Eddie AI can now produce material the shoot never captured. On the audio side, that means original music, voiceover in a voice you select, and dubbed tracks for an existing edit, with the mix handled automatically: music ducks under dialogue and speech is auto-leveled. On the picture side, you can describe a shot to search stock libraries and drop a playable candidate straight onto the timeline, or have the system generate b-roll outright.

All of it is optional and metered, and according to the company, none of it touches a cut unless you ask. Eddie AI also repeats a line that has defined its positioning so far: it does not invent your footage. Whether generated b-roll belongs in professional work is a question each production will answer differently, and anyone cutting for broadcast or brand clients should check how their delivery specs treat generative material before leaning on it.

Would you hand your rough cut to an assistant editor that lives inside a chat window, or does the edit still need to start in a timeline? Don’t hesitate to let us know in the comments below!