Attensira changelog

• Projects: a crew for each piece of ongoing work

  Agent

  Work is organised into projects, each a standing conversation with a crew staffed for one goal, such as winning the comparison questions on Perplexity.

  More

  You set the focus and answer what only you can; the crew posts what it did, and why, to the project's feed.

  Projects in the docs

• Goals

  Measurement

  A goal is one number from a fixed catalog, a target and a date.

  More

  Its starting point is the first reading taken after it is created, never a typed value; a number that cannot be read shows as unmeasured with the reason, and no goal sentence claims a number moved because of our work.

  Goals in the docs

• Markets: read prompts country by country

  Measurement

  Add a country, and optionally the language buyers there ask in, and every prompt is read again from that market with its own rates, n and trend.

  More

  Each engine says how it was located there, and one that cannot be read from a country shows as not measured.

  Markets in the docs

• Brand family

  Measurement

  Other domains, products and sub-brands you own count as you once you confirm them, in share of voice, citation rate and position.

  More

  Agents propose them with evidence; a confirmed property is never listed as a third-party source or a competitor.

  Brand family in the docs

• Google Search Console and rank tracking

  Measurement

  Search Console clicks, impressions and positions are kept day by day per page and query, and tracked Google searches are read daily with whether an AI Overview showed.

  More

  Both sit beside the AI answers, with the same rule that an unread number is a dash, not a zero.

  Google Search in the docs

• Page changes, measured before and after

  Measurement

  Each read of one of your pages is compared with the version on file.

  More

  A change, whether we shipped it, you made it from our advice or nobody suggested it, gets the same +7, +14 and +30 day re-reads as a merged pull request, without ever claiming the change caused anything.

  Page changes in the docs

• Google Ads Search campaigns, filed for approval

  Agent

  The agent researches keywords and drafts a Search campaign for your site.

  More

  Nothing reaches Google and nothing is spent until a signed-in person approves it; ad spend is billed by Google to your own account and never comes out of credits.

  Google Ads in the docs

• Shopify product and collection copy

  Publishing

  Connect a Shopify store and the agent rewrites product and collection copy and adds blog articles as unpublished drafts, instead of handing you text to paste.

  Shopify in the docs

Autopilot: the Daily win plan · Loop health · and 5 more

• Autopilot: the Daily win plan

  Agent

  Every workspace runs a Daily win plan that picks the next work toward the prompts you track, plus event automations that react when an answer or a page changes.

  More

  Everything they make waits for your approval or review; the loop never posts, sends or merges on its own.

  Autopilot in the docs

• Loop health

  Agent

  One view says whether the autopilot loop is working, idle or blocked, and why: readings, automation runs and skips, findings, work produced and re-reads due, per day.

  Loop health in the docs

• Client reports and the client reviewer link

  Agencies

  Weekly, monthly and quarterly client reports cover visibility as counts, work shipped and before-and-after verdicts, on a signed link you can forward.

  More

  A private reviewer link per workspace lets a client approve or decline suggested changes without an account, and see nothing else.

  Client reports in the docs

• Branding on shared pages

  Agencies

  Put your name, logo and colour on shared receipts, rewrites and client reports, set once for every workspace in the org.

  Branding in the docs

• Notion as your blog

  Publishing

  Connect a Notion database and approved articles arrive as draft rows; updates to existing posts wait for your yes, and nothing is ever deleted.

  Notion in the docs

• Community answers, drafted for you to post

  Agent

  When a Reddit, Hacker News, Stack Exchange or Quora thread decides an AI answer to a prompt you track, the agent drafts a reply for you to read, edit and post from your own account.

  More

  Nothing is ever posted for you.

  Community answers in the docs

• Webhooks

  Platform

  Signed HTTPS POSTs when a session finishes, work is filed, an article is written, a change ships, a report is ready, a scheduled run is skipped or a goal changes, so your own systems do not have to poll.

  Webhooks in the docs

Findings, held back when they are noise · Recall over the answers already read · and 1 more

• Findings, held back when they are noise

  Measurement

  Hourly rules watch your readings and file what changed.

  More

  Before a finding is filed, its last three days are tested against the seven before them; a change too small to tell from sampling noise is held and checked again an hour later.

  Findings in the docs

• Recall over the answers already read

  Agent

  Recall searches the stored answers for the passages nearest a question and returns them with the prompt, platform and day they came from.

  Recall in the docs

• The morning brief

  Agent

  One email a day at 07:00 in your time zone, and a longer note on Mondays: what shipped, early signals on the pages it changed, citations, goals and what the agents do next.

  Morning brief in the docs

Mention rate is called share of voice again

• Mention rate is called share of voice again

  Naming

  The analytics screens and the docs use share of voice for the rate of successful runs that name your brand, matching the API field.

  More

  It is still a rate over the same denominator for every brand, not a slice of a pie, so brands' rates do not sum to 100%.

  Share of voice in the docs

GitHub App integration · API keys

• GitHub App integration

  Agent

  The agent connects to a repository through a GitHub App rather than a personal token, and opens its changes as pull requests you review and merge.

  More

  This is the path for sites and docs whose content lives in a repo.

  GitHub in the docs

• API keys

  Platform

  Workspace API keys can be created and revoked from the account area, and authenticate requests to the public REST API.

  More

  Keys are scoped to the workspace that issued them.

  API keys in the docs

Inbox: approve or decline what the agent proposes · Auto top-up and custom top-up amounts · and 2 more

• Inbox: approve or decline what the agent proposes

  Agent

  The agent's proposed work now lands in a queue instead of executing silently.

  More

  Each item can be approved or declined, and the API exposes the same list, approve and decline operations the screen uses.

  Inbox in the docs

• Auto top-up and custom top-up amounts

  Billing

  Credit balances can be topped up by a custom amount, and a rule can refill the balance automatically when it falls below a threshold you set.

  More

  Refunds are recorded as a first-class ledger operation rather than an adjustment.

  Credits and top-ups in the docs

• Share of voice is now mention rate

  Naming

  The metric was renamed in the UI because "share of voice" implies a share of a fixed total, and this number is not one.

  More

  It is the proportion of successful runs in a window that named your brand, computed against the same denominator as every competitor's rate. The rates are independent and do not sum to 100%. The API field (ShareOfVoice) and the docs page (Share of voice) keep their names for compatibility.

  Superseded on 5 September 2026: Mention rate is called share of voice again

• AI crawler activity is available to the agent

  Agent

  Crawler-fetch data is now one of the tools the agent can call, so it can tell an access problem from a content problem: a page no AI crawler has fetched cannot be cited, and that is a different fix from rewriting it.

  AI traffic in the docs

Daily readings no longer cost credits · Keywords, groups and channels are now prompts, topics and platforms

• Daily readings no longer cost credits

  Billing

  Scheduled readings are covered by the plan's prompt slots and response budget.

  More

  Credits are spent on agent work, not on the measurement that runs on a schedule.

  Plans in the docs

• Keywords, groups and channels are now prompts, topics and platforms

  Naming

  The tracking vocabulary was renamed across the product to match what the objects actually are: you track prompts, group them into topics, and read the result per platform.

  More

  Old URLs and API fields were migrated in the same change.

  Prompts in the docs

AI traffic · Knowledge and skills · and 2 more

• AI traffic

  Measurement

  Server-side collection records which AI crawlers fetched which pages.

  More

  AI crawlers do not execute JavaScript, so a client-side beacon never saw them; the product generates the install for the server-side path.

  AI traffic in the docs

• Knowledge and skills

  Agent

  Facts about your business can be stored as knowledge and composed into the context the agent works from, so the same correction does not have to be repeated in every session.

  Knowledge and skills in the docs

• Public REST API

  Platform

  The same capabilities as the MCP server, over plain HTTP, authenticated with a workspace API key.

  The REST API in the docs

• Plans are measured in prompt slots and response budget

  Billing

  Plans are metered on how many prompts you track and how many model responses are collected against them, with sampling depth as the difference between tiers, rather than on a single opaque allowance.

  Plans in the docs