Juma
AI marketing agent that runs 700+ repeatable marketing Flows on live campaign data
Pricing
Free & Paid
Platforms
Web, Desktop app, Mobile...
Developer
Juma Labs (formerly Team-GPT)
Rating
4.5 / 5.0
Last Updated
September 29, 2026
Overview
Juma is an agentic marketing workspace rather than a chat assistant with marketing prompts bolted on.
The unit of work is a Flow - a saved, repeatable procedure that starts a clean thread, carries the finished procedure in its opening prompt and runs through to a deliverable: a case study, a LinkedIn carousel, a Google Ads waste audit, a campaign performance report, an ICP and persona set.
Read on juma.ai on 29 September 2026, the company states there are 700+ prebuilt Flows across analytics and reporting, content planning, content writing, research and briefing, advertising, and strategy, and a team can also describe its own.
Behind them sits a project knowledge base holding brand guidelines, documents, websites and instructions, so the same context is available to every run instead of being re-explained in each prompt.
The accounts doing the work are pooled at workspace level: Juma is billed on credits rather than seats, and adding a teammate changes nothing on the invoice while running more work does.
What separates the product from the crowded field of AI content tools is that its Flows have hands.
Read on juma.ai/integrations and juma.ai/flows on 29 September 2026, the platform connects to Google Ads, Meta Ads, LinkedIn Ads, Google Analytics (GA4), Google Search Console, HubSpot, Salesforce, Webflow, Shopify, Slack, Mailchimp, Stripe, Ahrefs, Notion, Google Drive, SharePoint, OneDrive, Canva, Figma, Instagram and Trello, and a multi-system Flow can pull campaign data, generate the report, send it to Slack and update HubSpot inside one run.
The same workspace can be driven from outside the app: Juma ships an MCP server so Claude, Cursor and other MCP clients can call it, and Enterprise customers can run it in a private cloud on AWS, Azure or GCP.
Compliance is published as SOC 2 Type II, ISO 27001, GDPR and HIPAA with zero data retention.
Pricing is metered by work done, and the vendor publishes the meter, which is rare in this category.
Read on juma.ai/credits on 29 September 2026, a credit meters tokens at a published weight per model - Claude Opus 5 at 714 credits per million input tokens and 3,571 per million output tokens, Claude Sonnet 5 at 429 and 2,143, GPT-5.6 Terra at 446 and 2,676 - and the page lists three runs measured in the company workspace in August 2026: a customer case study at 56 credits, a Google Ads waste audit at 192 and a cross-channel performance analysis at 364.
Failed runs consume nothing, plan credits are valid for the billing month plus the following one, top-up credits last 12 months, and the model weight table carries a last-changed date.
The trade-offs are structural.
There is no per-seat pricing to scale with headcount, but the credit pool is the real ceiling: the free plan is 300 credits per user, the smallest paid tier is $49 per month with 5,000 credits, and a heavy reporting and article month has to be estimated against the worked examples before committing.
Some scale claims are also inconsistent between the vendor own pages - the home page says 400+ marketing teams and shows a 4.9 G2 score while the pricing page says 250+ teams and the review block reads 4.8 from 32+ reviews - so treat those figures as vendor-surfaced rather than audited.
Use Cases
Running a full paid media report by pulling Google Ads, Meta Ads and GA4 data into one Flow and publishing the finished analysis
Producing SEO and GEO articles, content plans and internal link suggestions from a brief rather than from a blank editor
Generating social carousels and post packs for LinkedIn, Instagram and X from a single topic, adapted per platform
Building ICP and buyer persona documentation with pain points and messaging angles, then reusing it as project knowledge for every later Flow
Standardising repeatable campaign work - email campaigns, Google Ads copy, campaign concepts - so freelancers and in-house teams deliver from the same procedure
Who Is This For
In-house marketing teams whose bottlenecks are reporting and content execution rather than ideas
Agencies running several client projects that want shared context and unlimited seats instead of per-seat licences
Performance marketers who want Google Ads, Meta Ads and GA4 analysis without exporting data by hand
Content and SEO leads who need article and carousel production to follow one saved procedure
Marketing ops and AI leads evaluating an enterprise-compliant agent workspace with SSO, DPA and private cloud options
Key Features
700+ prebuilt marketing Flows
Saved procedures that run end to end and return a finished asset: SEO and GEO articles, LinkedIn and Instagram carousels, email campaigns, campaign concepts, Google Ads copy, ICP and buyer personas, competitor research and reports.
Flows that read and write to live accounts
Google Ads, Meta Ads, LinkedIn Ads, GA4, Search Console, HubSpot, Salesforce, Shopify, Webflow, Slack, Mailchimp, Stripe, Ahrefs and more - including connections that take real-world action such as publishing a page or updating a CRM record.
Every major model, one credit pool
Claude, ChatGPT and Gemini models plus image and research models run in the same workspace, with a published weight in credits per million tokens for each model and a dated rate table the vendor commits to updating.
Projects, knowledge base and Brand Hub
Brand guidelines, documents, web pages and instructions are stored once as persistent context per project and reused by every Flow, with custom agents, a shared prompt library and brand voice assets for the whole team.
MCP, API and enterprise controls
An MCP server exposes the workspace to Claude, Cursor and other MCP clients; Enterprise adds private cloud deployment on AWS, Azure or GCP, SSO and SAML, a DPA and an SLA on top of SOC 2 Type II, ISO 27001, GDPR and HIPAA compliance.
Pros & Cons
Pros
- Free plan includes every Flow, every integration and every model - the only difference between Free and Pro is the credit balance, and Pro credits roll over into the next billing month
- Published credit ledger: a dated table of credits per million input and output tokens per model (Claude Opus 5 at 714 and 3,571 against Gemini 3.1 Pro at 286 and 1,714 on 29 September 2026) plus three worked run examples, so spend can be estimated before buying
- Unlimited seats and workspace-pooled credits on every plan, so freelancers and agency partners can be invited without changing the invoice
- Flows act on connected systems rather than only drafting text - pulling Google Ads, Meta Ads and GA4 data, publishing to Webflow, posting to LinkedIn, sending to Slack or updating HubSpot
- Compliance stack published as SOC 2 Type II, ISO 27001, GDPR and HIPAA with zero data retention, which is what agency and enterprise buyers ask about first
- Failed Flows consume zero credits, and MCP support means the same workspace can be used from Claude, Cursor or a custom agent
Cons
- Credits are the binding constraint rather than seats: the smallest paid plan is $49 per month with 5,000 credits, and the vendor own worked examples range from 56 credits for a case study to 364 for a cross-channel performance analysis, so volume has to be modelled in advance
- Plan credits only carry into the following month - a light month is not banked for a heavy quarter, and top-up credits are the documented way to absorb a launch spike
- It is a workspace with 700+ repeatable procedures rather than a hands-off autonomous agent: audiences, offers and judgement still come from the team, and Flows need reviewing before they are trusted
- Scale claims differ between vendor pages (400+ marketing teams and a 4.9 G2 score on the home page against 250+ teams and 4.8 from 32+ reviews on the pricing page), so third-party ratings should be checked on G2 or Capterra
- Private cloud deployment, SSO and SAML sit behind Enterprise pricing, and the public pricing page quotes Enterprise as custom rather than fixed - budget a sales conversation if data residency is a requirement
Pricing Plans
Free
Full product access for evaluating Juma on real work, with every Flow, integration and model available.
- 300 free credits per user
- All 700+ Flows
- All integrations, including those that act on connected tools
- Unlimited seats
- Projects, knowledge base and prompt library
- All models
Pro Starter
For marketers and teams who want the work done rather than drafted, with a bigger monthly credit pool.
- 5,000 credits per month on the entry tier, with a slider up to 100,000 credits
- Credits roll over into the following billing month
- Top-up credit packs without changing plan
- Auto-recharge when the balance falls below a threshold you set
- Per-user and per-Flow usage reporting
- Everything in Free
Enterprise
For organisations that need dedicated infrastructure or contractual controls.
- Private cloud deployment on AWS, Azure or GCP
- SSO and SAML plus platform administration
- Security review support and DPA
- Custom contract and SLA
- Dedicated account manager and priority onboarding
- Invoicing and custom billing terms
Frequently Asked Questions
Editorial Review
Editorial Team
September 29, 2026
We read juma.ai/pricing and juma.ai/credits on 29 September 2026. The published numbers are unusually concrete for this category: 300 free credits per user, Pro at $49 per month with 5,000 credits on the entry tier and a slider to 100,000, unlimited seats on every plan, and a dated model weight table in credits per million input and output tokens (Claude Opus 5 at 714 and 3,571, Gemini 3.1 Pro at 286 and 1,714). Three worked run examples are also published, from 56 credits for a case study to 364 for a cross-channel performance analysis, and the page states that failed runs cost nothing.
Editorial Team
September 29, 2026
Editorial verdict: Juma is the most credible attempt we have reviewed at pricing a marketing AI agent on completed work rather than seats, and the Flow library plus live Google Ads, Meta Ads, GA4 and CRM connections is a genuine step beyond drafting copy. Two things to weigh before subscribing. First, modelling volume matters more than headcount - the credit pool is workspace-wide and plan credits only roll one month, so a launch-heavy quarter needs top-ups. Second, some vendor figures do not agree across pages (400+ marketing teams and a 4.9 G2 score on the home page against 250+ teams and 4.8 from 32+ reviews on the pricing page), so verify third-party ratings on G2 before treating them as evidence.
Related AI Agents
Have an agent to share?
Submit your AI marketing agent to reach thousands of marketers, agencies, and founders every month.
Submit Your Agent