InVideo AI Review 2026: Pricing, Credits & Agent Two

How much does InVideo AI cost in 2026?

InVideo AI is worth it when a production needs to invent footage, coordinate many current media models, mix generated and licensed stock assets, preserve creative context, and move from script or references into a directed edit. Current month-to-month prices are Plus $20 for 75 credits, Max $100 for 390, Generative $200 for 800, and Elite $1,000 for 4,250. Annual totals are $200, $1,000, $2,000, and $10,800, while credits are still delivered monthly and do not roll over. Basic Autopilot costs two credits per output minute, Pro 80, and Ultra 160 before actor charges; Agent Two and standalone models have variable action-level pricing. The main cautions are expiring balances, default always-allow generation, one-seat individual tiers, separate iStock quotas, model-specific output rights and limits, a separate InVideo Studio subscription, and the need to verify every generated shot, claim, likeness, caption, and public-sharing setting.

InVideo AI Pricing and Credits in 2026

These are the current USD web prices and included credits in InVideo's live pricing data, checked August 24, 2026. The pricing page opens on annual billing and rounds the monthly equivalent, so the amount charged for the year is the useful comparison. App-store pricing can differ because of taxes, currency conversion, and platform fees.

Plan credits reset on the renewal date and do not roll over. Extra credits bought separately last 12 months from purchase but also expire when the active subscription ends. Exports do not consume AI credits, although an export can consume the plan's iStock allowance. Model, resolution, duration, video input, audio, avatars, retries, and agent actions can all change the real cost.

InVideo at a Glance

What InVideo Actually Is

InVideo AI can start from an idea, exact script, URL, treatment, PDF, rough cut, reference image, video, or other production material. Agent Two reads those inputs against a shared Context, can develop scripts and storyboards, routes shots through generative or stock sources, adds voice and music, and carries decisions across the project. InVideo describes the interaction as an ongoing back-and-forth production rather than a one-shot template generator.

Agent Two divides control into layers. Free-flow agent chat handles broad direction. Playbooks store standing creative rules. Workflows lead a user through repeatable jobs such as casting, storyboards, localization, or campaign adaptations. Notebooks let the creator select a specific model, prompt and revise a single asset manually without breaking project context. This separation is useful: a user does not have to surrender every shot to automatic model routing.

Slate is the editing layer inside Agent Two. InVideo says it can line up approved takes, accept conversational changes, import footage, fonts, and jingles, and expose blade, transform, audio-level, LUT, color, transition, text, source-monitor, marker, and asset-bin controls. It supports Premiere Pro and DaVinci Resolve shortcut sets, four audio layers, and effectively unlimited video layers. That is a much stronger editing claim than older InVideo AI reviews reflect, although the current product is still designed around an agent-managed production rather than only raw-footage post-production.

Not every InVideo-branded feature belongs to the same subscription. InVideo Studio is the separate conventional template and manual timeline product. InVideo's own help center explicitly says AI and Studio require two subscriptions. Reviews that combine Studio templates, the Studio timeline, Agent Two, Autopilot, Magic Box, and standalone models into one undifferentiated checklist overstate what one purchase buys.

There are also two cost paths inside InVideo AI. Autopilot turns a brief or script into an assembled video at a published per-minute Basic, Pro, or Ultra rate. Agent Two and standalone models meter the reasoning and media created in each exchange. A five-second model clip, a one-minute stock explainer, and an agent that generates several candidate assets are therefore not comparable units even though all draw from credits.

The price cards require two conversions before they are comparable. Month-to-month costs are $20, $100, $200, and $1,000. The annual tab advertises rounded equivalents of $17, $85, $170, and $900, but charges $200, $1,000, $2,000, and $10,800 upfront. Over 12 months, those annual terms save $40 on Plus, $200 on Max, $400 on Generative, and $1,200 on Elite. They do not create an installment plan, and the included credits still arrive and expire on a monthly cadence rather than becoming one durable annual pool.

At month-to-month prices, the nominal plan-credit cost is about 26.7 cents on Plus, 25.6 cents on Max, 25 cents on Generative, and 23.5 cents on Elite. Annual prepayment lowers the effective cost to roughly 22.2, 21.4, 20.8, and 21.2 cents respectively, assuming every monthly credit is used before reset. Elite's smaller annual discount means Generative actually has the lower annual price per included credit. These ratios compare allowance, not output: different models can buy radically different seconds, resolutions, audio, or agent work.

Autopilot provides the cleanest capacity model. Plus's 75 credits can buy up to 37.5 Basic stock minutes, about 56 seconds of Pro, or about 28 seconds of Ultra. Max's 390 can buy 195 Basic minutes, 4.875 Pro minutes, or 2.4375 Ultra minutes. Generative's 800 can buy 400, ten, or five; Elite's 4,250 can buy 2,125, 53.125, or 26.5625. Human or Pro actors add 20 credits per minute, and these maxima ignore alternate takes, edits, model calls, audio, and other generation.

Basic, Pro, and Ultra describe media strategy as much as quality. Basic uses a compilation of licensed stock with no generated media. Pro generates every pixel with a collection of current models such as faster flagship variants. Ultra raises the quality ceiling with the most expensive versions, including 4K-capable video and upscaled imagery where supported. A factual explainer can be better served by controlled stock than synthetic footage; a product performance may require Pro; only a few hero shots may justify Ultra. Spending more is not automatically more truthful or more on-brand.

Agent Two uses Lite, Pro, and Ultra reasoning or generation modes in a more variable workflow. One response can reason about the brief, generate several images, produce video, design a voice, and call other media tools. The three-dot menu on a response shows the charge for that exchange; the project menu exposes cumulative Project Usage. Because the unit is the collection of actions the agent performed, two textually similar directions can cost different amounts. Preserve the prompt, permission mode, models, outputs, and usage record when evaluating repeatability.

The default permission mode is Always allow, which lets the agent generate directly. Users can switch to Ask before generating videos or Always ask before generating. An approval card shows the proposed prompt, model, duration, and aspect ratio and allows changes before confirmation. Project credit limits provide another containment layer. For any serious trial, change the default before the first broad instruction, set a project ceiling, draft cheaply, and promote only accepted shots to more expensive modes. Otherwise the agent can spend credits while exploring a direction the user never intended to fund.

Subscription credits reset on the renewal date and do not carry forward. Separately purchased top-ups last 12 months from purchase but also expire when the active subscription ends. That creates two clocks and makes cancellation a usage event, not merely a billing event. An annual subscriber still needs to consume each monthly allowance on schedule. Production managers should forecast by project month, retain a usage ledger, and buy a top-up only when its longer validity and cancellation risk are understood. Enterprise advertises different non-expiring-credit terms, which require a contract.

Standalone models are billed near the underlying provider's listed price and usually by output second for video. Resolution, input type, audio, duration, and model tier change the rate. InVideo's own current example contrasts a Kling 2.6 Standard generation around half a credit with a Veo 3.1 4K generation with audio at more than ten. The models and prices can change without a plan-name change. The setting panel and approval screen are the only reliable quote for a particular generation at a particular moment.

The Free plan is a live evaluation surface but not a stable capacity entitlement. It requires no card, exposes selected Agents & Models tasks, and currently resets its limited quota every Monday at 00:00 UTC. The official article does not promise one universal credit number. A free account can test image generation and animate an image into video, but watermark, model, duration, sharing, and availability should be read from the active interface. A weekly reset is not evidence that a production workload can be completed or monetized on Free.

The stock meter is separate from AI credits. Paid plans advertise unlimited standard stock from Storyblocks, Shutterstock, Pexels, Pixabay, and other included libraries, alongside a monthly iStock allowance of 100, 200, 1,000, or 5,000. iStock is deducted when the finished project is downloaded, not when the asset first appears in the draft. Re-exporting the same project with the same iStock asset does not consume it again, but using that asset in a different project does. An export paywall can therefore appear while AI credits remain.

iStock can be controlled before and after generation. The workflow settings accept a cap; a prompt can specify the desired count; Magic Box can replace or remove automatically selected iStock after the video exists. When iStock is reduced, the product can substitute Storyblocks or Shutterstock media. A manual iStock replacement is an exception: the current help page says Magic Box will not remove manually chosen assets. Audit the export panel, because a visual library logo in preview and a stock count at delivery represent different stages of the quota.

Agent Two's Context is the continuity layer. It is designed to hold scripts, references, character and world information, creative rules, uploaded media, rough cuts, decisions, approved takes, and other production knowledge. Multi-Input Intelligence interprets each upload against that project context rather than treating it as an isolated file. A PDF might be the client brief, a video the pacing reference, a product image the object that must remain stable, and a rough cut the sequence to improve. The agent still needs explicit source priority and rights; context is memory, not truth arbitration.

Expert Agents are specialized roles inside the production rather than separately purchased chatbots. The current Agent Two positioning describes a crew that can approach strategy, script, casting, world design, shots, sound, edit, and other creative tasks with domain-specific methods. That is useful when a broad brief crosses departments, but it also means one user message can trigger a cascade of decisions and paid outputs. Ask the system to propose a plan before generating, identify which expert is acting, and approve the irreversible media steps separately.

Playbooks store standing rules for how the agent should work: process, language, decision gates, option counts, and generation behavior. They can be attached across projects and run under Lite, Pro, or Ultra intelligence tiers. A good Playbook can encode brand vocabulary, evidence rules, review order, aspect ratios, prohibited imagery, model ceilings, and approval policy. It does not guarantee compliance. Generated shots, translations, products, logos, claims, and casting still need acceptance against the actual brand and legal source of truth.

Notebooks are the close-control layer. Nearly 200 models and utilities can be used model by model and prompt by prompt without leaving the project. Pages group assets by scene, sequence, or approval state. Threads assign addresses such as a scene and shot lineage that the agent can reference later. Direct annotation and regional edits let the user point at the problem rather than rewriting a global prompt. Upscalers, SDR-to-HDR conversion, audio stem separation, and other tools broaden the surface, but each transformation can create new quality, color, timing, or rights questions.

The handoff between free-flow agent work and Notebooks is the important architectural claim. A user can ask the crew to develop a scene, manually choose and refine a model output inside a Notebook, then return the approved asset to the agent without downloading and re-uploading it. That preserves lineage and context better than a loose folder of generations. Still, teams should export approved masters and record the source prompt, model, settings, consent, and license outside the project because product history is not a formal rights-management system.

Workflows turn repeatable jobs into guided runs. The nine launch workflows include casting, reference research, production design, storyboarding, ad key-visual variations, localization, product-detail-page assets, campaign transfer, and reframing. They ask questions, accept references, propose options, and pause for user choices rather than requiring a full bespoke prompt at every step. Custom procedures can be described, saved as markdown, and attached to later projects. This is strong procedural leverage, but it can scale an error just as efficiently as a good method.

The reframing Workflow illustrates credit-aware staging. It starts from 16:9 or other source shots, proposes how to extend them into a new canvas, generates still-image extensions first, and seeks approval before spending on video. That separates composition from motion and avoids funding a fully generated clip with the wrong framing. It remains synthetic outpainting: backgrounds, text, geometry, faces, products, reflections, and action outside the original frame are inventions that need a pixel-level review before the extension is animated.

Localization is broader than literal translation. Agent Two's Workflow asks which regional market is intended and can translate, transliterate, change currency, or culturally adapt casting and the creative. That is a meaningful production distinction: French for France, Quebec, and Francophone Africa is not one audience. It also increases review obligations. Pricing, offers, legal disclosures, cultural references, pronunciation, lip sync, text layout, captions, product availability, and licensed likenesses need a fluent local approver before publication.

Slate is the sequence editor inside Agent Two. It can ask the agent to line up approved takes against the script, reveal missing coverage, and accept conversational changes such as shorter versions, shot swaps, color notes, audio adjustments, or transition repairs. Manual controls include blade, transform, levels, LUTs, color, transitions, text supers, markers, a source monitor, asset bins, four audio layers, and effectively unlimited video layers. Premiere Pro and DaVinci Resolve shortcut presets reduce muscle-memory cost; they do not make Slate those NLEs or promise project interchange.

External footage, fonts, product windows, jingles, and other assets can enter Slate and be placed by the agent according to the script and Context. This supports hybrid production instead of forcing every pixel to be generated. The user still needs to check the 2GB per-file upload limit documented for the older AI editor, accepted image/video containers, codec behavior, font and music licenses, exact in/out points, color management, audio sync, and whether an Agent Two-specific path has different limits. One successful upload does not prove long-form post-production readiness.

The current upload help page lists MP4, MOV, WEBM, AVI, MKV, FLV, GIF, and other common video containers alongside images, with a 2GB limit per file. Large interview masters, multicamera sources, high-bitrate mezzanine codecs, separated production audio, and hour-long screen recordings can exceed that practical envelope even before generation begins. Transcode a test copy, preserve the original master, and validate duration, frame rate, audio channels, color, alpha, and sync. Do not assume a broad container list means every internal codec is decoded identically.

Autopilot and Magic Box remain relevant but should not be confused with Slate. Autopilot takes a prompt or exact script into a guided video assembly. Magic Box applies conversational changes to an already generated video, including media, stock, captions, music, voice, chapters, or duration. Older help documentation describes outputs from 15 seconds up to 20 minutes. Slate is the new agent-context timeline. InVideo Studio is the separate template and conventional editing product. A tutorial using one surface may not apply to the others.

Script-to-video has several intent modes. A user can instruct the system to use an exact script, ask it to generate a script from the idea, or start from a link and other reference. 'Use my script exactly' should be verified against the rendered voiceover and captions; automatic pacing, scene division, pronunciation, timing, media selection, and truncation can still change the communicated result. For regulated, scientific, financial, medical, or client-approved copy, compare every spoken word and on-screen claim with the signed source.

The older AI editing surface provides useful manual repairs: add, remove, replace, crop, fit, rotate, zoom, and retime media; add or delete chapters; edit script and on-screen text; style or move subtitles; change transitions; replace music; change voiceover; adjust voice speed or delay; and add custom sound effects. Those operations make the output repairable. They are not evidence of full source-timeline parity, multicamera sync, advanced compositing, nested sequences, collaborative locking, color-managed finishing, or deterministic offline recovery.

Exporting does not use AI credits, but delivery still has entitlements. Paid plans promise unlimited watermark-free exports. Preview pages can display library, InVideo, or username watermarks that disappear only on an eligible export with branding disabled. A project exported on Free keeps its watermark after an upgrade until the original project is re-exported. If a paid subscription ends, previously generated media remains accessible, but editing an old paid project and downloading the changed version without a watermark requires the paid plan again.

Sharing visibility is a distinct privacy control. Current help says account videos are not visible to others by default, but when a user chooses to share, the link defaults to public unless the Default export visibility setting is changed. Private limits preview to the owner or workspace; unlisted still requires a logged-in InVideo AI user; public lets anyone with the link preview. Before sharing confidential work, set the account default, inspect the individual export, test the link from a signed-out browser, and revoke any obsolete public version.

Generated media can be downloaded independently of the full video and remains accessible after subscription cancellation under current help guidance. That is valuable for taking a hero shot into another editor. It should not become the only archive: download the selected output, retain prompt and settings, record the model and provider, save consent evidence, and preserve the applicable stock or model terms. Availability after cancellation is a product statement, not a guarantee that the same generator, price, model version, or regeneration settings will remain.

4K belongs to particular paths. Ultra raises the ceiling, and the current Veo 3.1 documentation supports four-to-eight-second generations, 720p through 4K, 16:9 or 9:16, text or one-to-three images, optional first and last frames, and native synchronized audio. Other models may top out at 1080p, omit audio, accept different ratios, or restrict faces and reference inputs. A plan's access to a model catalog is not a guarantee that a whole multi-shot export is natively 4K or visually consistent.

First and last frames, character references, Context, and consistent prompts help continuity, but each generated shot is still probabilistic. Veo's first/last-frame control can bridge planned endpoints; it cannot guarantee the path between them, exact identity, product geometry, text, hands, wardrobe, lighting, spatial logic, or lip movement. Multi-shot productions need a continuity bible, shot IDs, accepted reference assets, model and seed records where available, and a full sequence review. A good isolated frame can still cut badly beside its neighbors.

Avatar choices include the user's own Free or Pro clone, licensed human actors based on real people, and synthetic virtual actors. Creating a Pro avatar currently costs 20 credits, and Human or Pro library actors add 20 credits per output minute in standard plan documentation. Multiple avatars can adjust the total, so the approval screen is authoritative. The user should separate the cost of creating a clone, storing it under a plan quota, and using it in a video.

Voice cloning accepts MP3, WAV, or M4A and requires a permission statement at the start of the recording plus explicit checkboxes. The help center describes these controls through an older AI Twins flow, which is reached from Older versions in the latest interface. Consent is necessary but not sufficient: the publisher must also control who can use the clone, which claims it can deliver, what happens after cancellation, how impersonation is prevented, and whether local biometric, employment, advertising, or union rules apply.

InVideo prohibits creating a celebrity avatar without authorization and publishes blacklisted topics and use restrictions for avatar actors. Human actor libraries compensate the person whose likeness is used, which explains some incremental charges. Current help permits monetization of avatar videos, but commercial use does not waive defamation, false endorsement, disclosure, publicity, platform, or sector-specific obligations. Review every script and context in which a real or synthetic persona appears, and preserve the approval record for a client's spokesperson.

The platform says it can generate videos in more than 50 languages and identifies a smaller group as fully tested. Subtitle support is narrower, with current gaps including Arabic, Urdu, Hebrew, Simplified Chinese, Tamil, Marathi, Maori, Farsi, Thai, and Hindi. Generation language, voice availability, pronunciation, caption support, text layout, localization workflow, and right-to-left rendering are separate capabilities. Test the complete source-to-export path with a fluent reviewer rather than assuming a translated script guarantees correct captions.

Folders & Connect are Team and Enterprise controls. Internal or external collaborators can be invited into a specific folder instead of the whole workspace, and outside partners do not necessarily need a paid seat. Work remains inside the organization's project boundary, with budget controls available by user or project. Individual Plus, Max, Generative, and Elite subscriptions are one-seat products. A solo plan's large storage or credit allowance should never be priced as if it includes an agency review team.

Enterprise marketing promises unlimited seats, non-expiring credits, per-project and per-user budgets, ownership of inputs and outputs, no model training on customer data, SOC 2 and GDPR alignment, and one contractual relationship across more than 200 model providers. Those are procurement claims, not self-serve entitlements. Buyers should request the current security report, DPA, subprocessors, data locations, deletion SLA, incident terms, IP indemnity scope, model-provider pass-through terms, support response, audit logs, SSO, retention settings, and the exact meaning of ownership for non-unique AI output.

The beta remote MCP server is documented at https://mcp.invideo.io/sse for Claude.ai, Claude Desktop, Cursor, Windsurf, and compatible hosts. InVideo's help page recommends the hosted endpoint because it requires no authorization. That lowers setup friction and raises governance questions: the 2025 article does not document a current tool catalog, identity scope, rate limits, credit linkage, output persistence, or the relationship to Agent Two. Test it in a non-confidential sandbox and do not assume an unauthenticated transport has access to a paid workspace.

InVideo AI and InVideo Studio still require separate subscriptions. Agent Two's Slate narrows the feature gap by providing a real sequence editor, but Studio retains its own template catalog and conventional editing workflow. A comparison page that says 'InVideo has templates' or 'InVideo has a timeline' without naming the product and subscription is not purchase-grade evidence. Decide whether the job needs generated production, Agent Two context, Studio templates, or both, then price each checkout and test asset movement between them.

Cancellation leaves the subscription active through the paid term, after which the workspace moves to Free. Monthly plan credits and separately purchased credits do not survive indefinitely; paid-project revisions may regain a watermark restriction; some generated media remains downloadable. Refunds are much narrower: InVideo says a request within seven days of purchase or renewal may be considered only if no video or asset was generated or downloaded in that billing period. A trial generation can therefore close the refund path immediately.

Workspace deletion is different from deleting a project. The current account procedure goes through Settings, Manage account, Delete workspace, and Delete account, with a support form if the self-serve path fails. Export approved videos, generated media, prompts, scripts, consent evidence, usage records, invoices, and licenses first. Remove collaborators, public links, and integrations. Confirm subscription cancellation and retention requirements. Deletion is an access and data event, not a billing reconciliation or guaranteed recovery mechanism.

The Terms grant broad rights to use generated output, state that InVideo will not train or fine-tune its AI/ML models on User Data, and permit de-identified or aggregated information for analytics, quality, security, and improvement. Users must own or license every input, and model or stock assets can carry separate conditions. AI output may not be unique. Publicity, privacy, copyright, trademark, music, stock, actor, disclosure, and provider restrictions remain with the publisher even when the final export is commercially usable.

The most reliable evaluation is a production ledger. For one real brief, record plan, permission mode, every response and model charge, top-ups, Basic/Pro/Ultra mode, iStock selected and consumed, actor fees, generation attempts, accepted assets, human correction time, subtitle errors, continuity failures, Slate repair time, export settings, watermark status, public-link visibility, and final approved duration. Divide total cash and labor by approved deliverables. Compare the same footage-grounded job in Valmera and the same finishing job in the team's NLE.

The decision boundary is straightforward. InVideo is strongest when the project needs to create what was never filmed: a campaign world, synthetic performance, storyboard, localization set, product ad, generated music, or model-orchestrated sequence. Valmera is stronger when the uploaded recording is primary and the job is to cut, clean, reframe, caption, score, and revise that real source through conversation. An NLE is stronger when the timeline, interchange, color, sound, and recovery must remain deterministic under a human editor's direct control.

Where InVideo Wins

Limits and Trade-offs

Who Should Choose InVideo?

• Marketers creating explainers, performance ads, UGC-style ads, social campaigns, and visual stories from scripts, links, or briefs.
• Filmmakers using AI for casting studies, storyboards, previsualization, synthetic shots, rough assemblies, or end-to-end generative productions.
• Creators who want many current image, video, voice, music, and utility models without opening a separate account for every provider.
• Teams that need shared project context, repeatable Playbooks or Workflows, project-scoped collaborators, and generation-budget controls.
• Publishers who need licensed stock, voiceover, music, captions, avatars, and generated media without beginning with a large footage library.
• Hybrid productions that can reserve expensive generation for missing or impossible shots while using stock and uploaded media for the rest.
• Creative directors who want an agent to preserve a campaign bible across strategy, casting, world building, shots, sound, and edit.
• Teams that can formalize brand rules and approval gates in Playbooks, then run repeatable jobs through guided Workflows.
• Creators who want close manual control over one generation in Notebooks without losing the wider project's Context.
• Editors willing to validate a new agent-directed sequence workflow in Slate while retaining a conventional NLE for deterministic finishing.
• MCP users who want to experiment with an official beta remote endpoint in a non-confidential, non-production sandbox.
• Enterprises that need folder-scoped collaborators, per-user or project budgets, negotiated credits, security review, and a single model-provider contracting layer.
• Production teams that can mix inexpensive stock assembly with selectively funded flagship 4K or native-audio hero shots.
• Multimarket advertisers with fluent local reviewers for regional translation, currency, casting, captions, claims, and cultural adaptation.

InVideo vs Valmera

InVideo AI and Valmera both use an agent, but the source of truth differs. InVideo is strongest when the source is an idea, script, production bible, references, and a choice of generated or stock assets. Valmera is strongest when the source is irreplaceable footage and the task is post-production: understand it, select it, cut it, clean it, caption it, reframe it, augment it, preview it, and render it.

Agent Two's Slate narrows the gap because InVideo can now assemble approved shots, import external media, and accept conversational precision edits. The deciding test is still the normal job. Give InVideo one real production brief with the expected model quality and retries; give Valmera one representative raw-footage edit with the expected revision loop. Compare the accepted final output, not the first demo.

Pricing should be normalized by deliverable. InVideo credits can represent agent reasoning, images, generated seconds, audio, actors, or Autopilot minutes and disappear at renewal. Valmera charges editing-agent turns from its own combined balance. Neither vendor's credit equals the other's. Measure total credits consumed across draft, failed generation, revision, export, and human cleanup.

Choose InVideo when you need to invent footage, span many frontier models, create avatars or synthetic ads, use premium stock, or manage a generative film through Context, Notebooks, and Slate. Choose Valmera when preserving the meaning and continuity of recorded interviews, events, multicamera sessions, narrative footage, or a creator's own takes is the non-negotiable requirement.

Agent Two and Slate make InVideo more editor-like than its older prompt-to-stock reputation suggests. The center of gravity is still production creation: Context, expert agents, synthetic media, stock, avatars, model routing, and campaign workflows. Valmera's center is transforming a user's uploaded recording from an open-ended brief. Both use conversation, but they attach it to different source truth and cost structures.

A fair cost test separates three jobs. For a generated one-minute ad, price InVideo's mode, actors, models, hooks, retries, stock, and human review. For a real interview edit, price Valmera's turns and rendered revisions against the same source. For deterministic finishing, price editor hours in Premiere or Resolve. One nominal AI credit or one successful render cannot be compared across those jobs.

InVideo's MCP is an external entry point into an AI creation brand; Valmera's MCP and app surfaces are designed around project-scoped video-editing operations. Tool availability, authorization, persistence, and current tool catalogs should be verified directly before either is selected for automation. The marketing phrase 'MCP video editor' does not establish equivalent control or safety semantics.

For the full feature-by-feature comparison, read Valmera vs InVideo.

Primary Sources Checked

• InVideo AI — Pricing
• InVideo Help — Plans and Credits
• InVideo Help — Free Plan
• InVideo Help — Basic, Pro, and Ultra Costs
• InVideo Help — Agent Credit Charges and Approval Controls
• InVideo Help — AI Filmmaking Overview
• InVideo Help — Agents and Models
• InVideo — Agent Two
• InVideo News — Agent Two Notebooks
• InVideo News — Agent Two Workflows
• InVideo News — Agent Two Playbooks
• InVideo News — Agent Two Multi-Input Intelligence
• InVideo News — Slate Editor
• InVideo News — Agent Two Stock
• InVideo News — Folders and Connect
• InVideo News — Enterprise
• InVideo Help — AI vs Studio
• InVideo Help — Magic Box Commands and Duration
• InVideo Help — Upload Formats and 2GB Limit
• InVideo Help — Script-to-Video Modes
• InVideo Help — Supported Languages and Subtitle Gaps
• InVideo Help — Team Members and Seats
• InVideo Help — iStock Consumption
• InVideo Help — Avatar Types
• InVideo Help — Voice Cloning and Permission
• InVideo Help — Veo 3.1 Specs and 4K
• InVideo Help — Dynamic Text and Captions
• InVideo Help — Downloading Generated Media
• InVideo Help — Cancellation
• InVideo Help — Refund Eligibility
• InVideo — Terms, Output Rights, and Data Training
• InVideo Help — Remote MCP Server
• InVideo Help — Current Usage
• InVideo Help — Credit and iStock Rollover
• InVideo Help — Generation and Export Paywalls
• InVideo Help — Pre-Generation Credit Display
• InVideo — Video Model Credit Pricing
• InVideo — Image Model Credit Pricing
• InVideo News — Agent Intelligence
• InVideo News — Expert Agents
• InVideo Help — Workspace Deletion
• InVideo Help — Paid Projects After Downgrade
• InVideo Help — Downloading Finished Videos
• InVideo Help — Export Visibility
• InVideo Help — Paid-Plan Watermarks
• InVideo Help — Export Username Branding
• InVideo Help — Create a Personal Avatar
• InVideo Help — Library Actors
• InVideo Help — Avatar Monetization
• InVideo Help — Avatar Access After Cancellation
• InVideo Help — Multiple Avatars
• InVideo Help — Avatar Content Restrictions
• InVideo Help — Celebrity Avatar Restriction
• InVideo Help — Autopilot Tutorial
• InVideo Help — Video Aspect Ratio
• InVideo Help — Subtitle Styling
• InVideo Help — Previously Generated Media

Frequently Asked Questions

Want the AI to Perform the Edit?

Upload your footage to Valmera, describe the outcome, and refine the rendered edit through conversation.