Run Google Ads and OpenAI Ads with AI agents.
Adjar is the control layer for AI agents across your ad accounts. Your account becomes plain-text config: the agent drafts a change through the adjar CLI, you review the diff, Adjar applies it safely. The next change starts from the performance report, read beside the config.
# account root — composes the restplatform = "google"imports = [ "google/assets.toml", "google/conversions.toml", "google/campaigns/*.toml",][account]id = "4821937650"sitelinks = [ "sl-size-guide", "sl-returns" ]callouts = [ "co-free-shipping", "co-reviews" ]● Read campaigns.toml
● Edited campaigns.toml
3 keywords status ENABLED → "PAUSED"
1 campaign daily_budget_usd 200 → 240
Run `adjar plan` to review the diff before anything goes live.Engineer the loop — it's all plain text.
Your config and the performance report are both plain text. The agent drafts the next change, you review the diff and apply it, and the report becomes the next ask. For ad accounts, that loop is Adjar.
Who Adjar is for.
A growth team of one, an agency book of clients, a company-wide AI push — the same review-and-apply workflow.
Run paid media as a team of one
Own every campaign alongside the rest of growth? Put your AI agent to work: daily reports land next to the config, the agent spots waste and drafts the fix, you review the diff. Tuning that took days ships in minutes.
- Ship campaign changes as reviewable diffs, not console clicks
- The agent reads daily reports and proposes fixes itself
- Every change has an author, a review, and a way back
| Campaign | CTR | Cost | Conv. |
|---|---|---|---|
| Brand | 7.4% | $38.20 | 26 |
| Trail Running | 2.9% | $182.40 | 9 |
| Trail Running · DE | 2.1% | $41.13 | 2 |
The trail shoe group is catching "shoe repair" queries — adding an ad-group negative:
+ [[campaigns.ad_groups.negative_keywords]]
+ text = "repair"
+ match_type = "BROAD"Your ad account, as a file your AI can run.
The console was built for human clicks. As plain text, your account does three things it never could.
Your AI agent can actually use it
Agents read and edit text reliably; a console of clicks defeats them. Plain text is what makes an AI agent useful on your ads at all.
It scales past where clicking gives out
Change 200 campaigns with one find-and-replace, not 200 rounds of clicking — and clone what works into a new market in seconds.
Every change is tracked and reversible
Every change is a saved revision, so your account keeps a version history, like code. Review the diff, then revert if one goes wrong.
Your agents run everything but the ads — and those still mean a thousand clicks by hand.
Google Ads and OpenAI Ads today — more platforms coming.