Share the page and the target query, and Juma writes meta tags that fit the character limits and out-compete the snippet currently winning that SERP slot.
Share the page URL and the target query. Juma pulls the live SERP for that query, identifies which competitor's snippet is winning, and writes meta titles capped at 60 characters and meta descriptions capped at 155 characters that directly compete against the current winner. The Flow returns paste-ready strings for the CMS, the rationale behind each angle, the live competitor snippet shown side-by-side for comparison, and a current-state diagnostic flagging any meta field that is missing, length-exceeded, or falling back to body copy.
Name the brand, the category, and the page count, and Juma identifies the URLs, pulls the live SERP for each target query, extracts the snippet patterns currently winning each slot, and writes a fresh meta title and description per page calibrated against the SERP it is competing in. The output is an XLSX workbook with one row per URL covering current title, current description, proposed title, proposed description, character counts for each, the SERP winner for context, and a status column the CMS team uses to track implementation. The workbook is sortable and filterable, with conditional formatting that flags any field outside the 30 to 60 character title range or the 120 to 155 character description range.
Generate fresh meta titles and meta descriptions for 20 of Notion's project management category pages. For each page, pull the live SERP for its target query and write descriptions that differentiate from what is currently ranking. Deliver as an XLSX spreadsheet.
FAQ pages, support docs, and "what is" content often rank but lose featured snippets to competitors because the meta description is generic body copy instead of a direct answer. Juma scrapes each page, checks which SERP feature it currently triggers (AI Overview, PAA, featured snippet owner), and rewrites the meta description in featured-snippet format: definition-style for "what is" queries, step opener for "how to" queries, list opener for "best of" queries. The meta title aligns to the exact target question so Google's featured-snippet algorithm picks the page up. The output is an Excel workbook with tabs for Summary, Rewrites with current vs. new for every page, character counts, SERP feature notes, status tracking, and Sources.
Notion has 30 FAQ pages getting impressions but no featured snippets. Rewrite their meta titles and descriptions so each description doubles as featured-snippet bait for the page's target query.
Name the target query and the brand, and Juma writes a set of variants where each one takes a different angle: a free-pricing or feature-bundling counter to the current winner, a direct "only" claim counter that neutralizes a competitor's exclusivity language, a specificity cut-through for high-intent buyers tired of listicle results, or a social-proof variant leading with G2, Capterra, or named customer signal. Each variant comes with character counts, the angle explained in one line, and the competitor snippet it was designed to compete against. The team picks the angle that fits the brand's voice. The variants are paste-ready into the CMS without further editing.
Write 5 meta title and description variants for Notion's "Project Management" landing page, each taking a different positioning angle against Asana, Monday, ClickUp, and Wrike. Show the competitor snippet each variant is responding to.
Share a sitemap or a list of URLs, and Juma scrapes every page, extracts the current meta title and meta description, runs the character-length validator against the 30 to 60 character title range and the 120 to 155 character description range, and flags every field that is missing, empty, length-exceeded, or falling back to body copy. The output is an XLSX with a row per URL, a column for each issue type, and a priority column based on whether the page has organic traffic to lose. Pair the audit with a follow-up rewrite prompt for the worst offenders so the fixes ship in the same session.
Audit the meta titles and descriptions across the top 50 pages on https://www.notion.so. Flag every field that is missing, over the character limit, or falling back to body copy. Deliver as an XLSX with priority based on organic traffic.
Meta tags do double duty when they mirror the same target query the page is optimized for under generative engine optimization and answer engine optimization. Juma pulls the GEO audit's primary target query and the AEO answer block from the same page, then writes a meta title that echoes the H1 pattern, a meta description that previews the page's direct-answer 40 to 60 word opening, and FAQ schema-aware language that helps Google connect the meta description to the FAQPage markup. The result is meta tags that work for traditional SERP, featured snippets, and AI Overview source attribution simultaneously.
Take the GEO and AEO findings from Notion's project management page and write meta titles and descriptions that align with the on-page direct-answer opening and FAQ schema. The meta description should preview the answer block, not summarize the page.
A Juma Project is a shared space where the team stores everything Juma needs to know about a client's meta-tag program. Create one project per client, add context as the program develops, and Juma uses what is relevant every time the team runs a flow. The more the team adds, the sharper every variant and every bulk refresh gets.
Target Query List
The exact SERP queries each page is competing for, written as a per-URL mapping. "https://notion.so/product/projects → project management software" not "project management." With the list in the project, every bulk run uses the correct competitor SERP per page instead of guessing the query from the URL.
Page Set
The 20 to 100 URLs the team most wants to keep fresh. Usually the homepage, the category pages, the comparison pages, and the FAQ pages. Pair each URL with the target query above. Juma reuses the list every refresh, so trend tracking is consistent month over month.
Character Standards
The brand's house standards for title length, description length, and any banned words or required brand-name suffixes. Some brands insist on the pipe character; some prohibit ellipses. With the standards in the project, every generated variant respects them without the team having to flag exceptions.
Brand Voice Guide
Tone, vocabulary, banned words, and approved phrasings. Juma applies the voice to every meta title and description so the team is never publishing on-brand pages with off-brand SERP snippets.
Add a short description to each knowledge item in the project's info field so Juma knows what each file contains and when to use it. For example:
Meta titles should land between 30 and 60 characters, with the strongest performers hitting 50 to 58 characters so the full title displays on both desktop and mobile SERPs without truncation. Meta descriptions should land between 120 and 155 characters, with the strongest performers around 150 to 155 characters so the description fills the snippet box without spilling over the cutoff.
Google occasionally renders longer descriptions on desktop, but the team should not write to that range because mobile and AI Overview snippet boxes still enforce the lower bound. Juma's `analyze_geo_technical` tool validates every generated meta tag against the 30 to 60 / 120 to 155 ranges and flags both over-length and under-length fields. Pixel-width estimators add a second layer for brand suffixes that take more visual space than their character count suggests.
Most meta description generators output a generic summary based on the page title or a paste of body copy. Juma pulls the live SERP for the page's target query, identifies which competitor's snippet is winning, and writes the meta description to compete against that specific snippet. Every variant comes with the competitor snippet shown side-by-side so the team sees what each angle is positioning against.
The bulk version goes further: Juma runs `analyse_google_serp` per URL, extracts the snippet patterns that dominate the category, and writes differentiated descriptions so the brand does not echo the same generic structure 20 pages in a row. The output is an XLSX workbook the CMS team can implement page by page, with character counts and status tracking already wired in.
The workbook is a paste-ready XLSX, not a direct CMS integration. Most CMS tools accept bulk meta-tag uploads via CSV import or a developer-facing API endpoint. The workbook columns map cleanly to the standard fields most platforms expect: URL, current title, current description, proposed title, proposed description, character counts, and a status column for tracking which rows have shipped.
For Search Console specifically, Juma can pull GSC CTR data per URL when the integration is connected on the workspace, then prioritize the bulk refresh by the pages with the lowest CTR despite high impressions. That ordering surfaces the highest-leverage refresh candidates first instead of running a flat sitemap audit.
Yes. The XLSX workbook ships with a status-tracking column that records the date each row was refreshed and the team member who shipped it. Saved to the project, the workbook becomes the source of truth across months: the next refresh inherits the prior status column, so the team can filter to "not refreshed in 90 days" or "refreshed but CTR did not move" cohorts.
The trend view also catches when a competitor's snippet changes mid-cycle. If Asana rewrites its description, the next bulk run will surface the new competitor snippet in the SERP-winner column, and Juma flags the rows where the brand's existing description is now writing against an outdated competitor pattern.
Yes. The character math is universal across categories. What changes is the SERP landscape Juma writes against. B2B SaaS pages compete against G2-style review listicles and direct competitor product pages. Ecommerce pages compete against Amazon, Wirecutter, and category-comparison content. Agency brands compete against G2, Clutch, and trade publications.
The Flow asks for the target query early in the session so every variant reflects how the audience actually searches. The character standards and character-length validator are universal. The competitor-snippet research is category-specific. Strategy and the call on which angle to ship stay with a senior content lead. Juma handles the SERP research, the variants, and the workbook.
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
Start typing a task and Juma will figure out the rest.
