POWERED BY FEATHER · LIVING CONTEXT ·

Dashboards, decks, weekly reviews. That era is over.

The Performance Agent.

An AI media buyer for performance marketing. Runs Meta, Google, and YouTube against your KPI, 24/7.

Avg +25% ROAS in 90 days 30-day pilot, free Outcome-priced Data never leaves
app.hawky.ai/d2c-grooming-in/team/PERF/active
Active experiments PERF
ROAS 30d
4.1× ▲ +1.6× vs prior
Creative Health 30d
82/100 ▲ +14 after refresh
CTR 30d
2.84% ▲ +1.04pt vs prior
Status:Active, In review × Add filter
In Progress 6 +
PERF-241 Bid +15% AM on top-3 creatives — CPM softened −8% off-peak bidding M 2m
PERF-239 Hook test: "before/after" vs "day-in-life" · 6s pre-roll creativecopy Y 9m
PERF-237 Reallocate $8.2k → Lookalike +3% · ROAS impact +0.4× projected budgetaudience M 14m
PERF-235 Interest stack v2 — male 25–34, gym + skincare overlap audience M 22m
In Review · awaiting human 3 +
PERF-243 Approve creative refresh — male 35–50 · brand-safe lock creative M SR 3m
PERF-242 Budget shift: scale Meta +22% at the cost of YouTube — confirm? budget M KP 12m
Done 38 · this cycle +
Trusted by performance teams at
TVS Motors Cars24 Swiggy Instamart Snapdeal Matrimony.com Titan Bombay Shaving Company Hiveminds Razorpay Royal Enfield Jawa Noise La Marquise smallcase Scaler Bhanzu GUVI Blink Digital Yellow.ai Livspace SolarSquare
Dispatches · from the pilot fleet

What performance marketing is going to look like.

Three operators ran scoped pilots: one objective, one region, one goal each. Same conclusion across all three. Read these as field reports, not case studies.

Transmission 01 Verified · live
0:42
Pramod
Head of Growth · Bhanzu
PILOT 01 EdTech · India Single-objective scope

We piloted it on one objective: re-engaging Hindi-speaking parents who had churned. This is what performance marketing is going to look like in two years. Maybe one.

Objective
Retention re-engagement
Region
Hindi-speaking · India
Goal
Reactivate dormant cohorts
Transmission 02 Verified · live
0:51
Nandani
Head of Growth · Instamart
PILOT 02 Q-Commerce · India Three-city scope

We pointed it at one category in three cities. The shift wasn't in the dashboards. It was watching the agent move at the speed our category actually moves. The old playbook is over.

Objective
Hyperlocal acquisition
Region
3 metros · staggered launch
Goal
New-user onboarding velocity
Transmission 03 Verified · live
0:38
Jayant
Head of Growth · Bombay Shaving Co.
PILOT 03 D2C Grooming · India Single-campaign scope

We ran it on one campaign: first-time-buyer acquisition. The realization came on day three. This is the operating model for performance marketing from here on. Everything before this is the manual era.

Objective
First-time-buyer acquisition
Region
Tier-1 metros · India
Goal
Cold-traffic conversion

Your ads run 24 hours.
Even the best teams need sleep.

Strategy, brand, and creative direction belong to your team. The 16 hours after they sign off belong to whoever's still on shift. The agent handles bid moves, budget shifts, and fatigue catches in those hours, so your team stays focused on what only humans can do.

The work that compounds while you sleep.

Auctions, fatigue, and overnight drift don't keep office hours. The agent runs through the night so your team can lead the morning.

Yesterday's leak, found this morning.

By the time the dashboard opens, the wasted spend is already gone. Detection-after-the-fact is the only mode the old stack supports.

70% of your buyer's week on routine ops.

Budget shifts, pause/resume, bid tweaks. Work that fills the calendar but earns nothing strategic.

70%
ROUTINE OPS STRATEGIC

Every campaign starts from zero.

What worked last quarter is fragmented across Meta, Google, CRM, and Slack. Nothing compounds. Every launch is a cold start.

The work the agent does, while you do the strategy.

Plans. Launches. Optimizes against your KPI. Mines your category for what's working. Logs every move with the data that triggered it, and lets you roll any of it back.

Deploying · live

Plans, launches, and runs your accounts.

The agent ships campaigns, ad sets, and creatives straight into Meta, Google, and YouTube, and keeps them running. No drafts. No approvals queue. No waiting around.

Shipping today 147 ads · 12 in queue · 00:00→now
23:48 AD-1247 Hero v3 · Reels · male 25–34 · IN Live
23:46 AD-1246 PMax · "best for sensitive skin" · 4 assets Live
23:45 AD-1245 Demand Gen · 6s pre-roll · MENA · 3 cuts Live
23:43 AD-1244 Reels · UGC v2 · 9:16 · LAL 1–3% Live
23:41 AD-1243 Story · "before/after" · LAL 1% · IN tier-1 Shipping
+ 137 earlier today 89 Meta 42 Google 16 YouTube
Optimizing · KPI-locked

Optimizes against your KPI, not engagement.

You set the goal: ROAS, CAC, LTV, or contribution margin. The agent allocates, scales, and kills against that goal. Nothing else gets a vote.

Q2 Goal
CAC$24 · ROAS3.5×
Current 4.1× +17% vs goal
Suggestions 3
Scale Hero-v3 · +51% budget
CAC −$3.10 · confidence 0.93
Shift Meta → PMax · $4.2k
ROAS +0.4× · confidence 0.91
Kill Age 45+ · CAC $48 · under bar
save $48/d · confidence 0.96
Sensing · category & competitors

Mines what's working in your category.

Watches competitor creatives, hook patterns, claims, pricing moves, and emerging keywords. Tells you what to copy, what to avoid, and what to ship next.

Mining: skincare · D2C · IN
Live
Insights 4 new
Hook “before/after” UGC pattern
+63% reach · n=412 · c=0.91
!
Claim “under $30” beats “premium”
CPM −22% · category · n=412
Search “best for sensitive skin”
+63% WoW · Google Trends · geo IN
Competitor Acme · 14 new UGC hooks
3 outperformed your hero-v2 · 24h
Audit log · 310 today

Logs every move. Reversible.

Every decision ships with the data that triggered it, the confidence score, and a rollback button. Set guardrails on what the agent can touch: budget, creative, geo. You stay in command.

23:48 Hero-v3 budget $8.2k → $12.4k (CAC−14%, c=0.93) SCALE
22:11 Bid AM +15% on top-3 creatives (CPM−8%, c=0.91) REBID
19:34 Paused Age 45+ set (CAC $48 · over bar 14d) PAUSE
17:02 Rotated Hero-v2 → Hero-v3 on saturation (idx 0.78) ROTATE
+ 306 earlier today Roll back any
See it run on one of your accounts. Book a demo

Every dollar makes the next dollar smarter.

A closed loop that runs while you sleep. It's a system, not a report. Four steps that repeat every minute the auction is open, each one feeding the one after it. Powered by Feather's Living Context.

FeatherDB LIVING CONTEXT SHARED MEMORY TAGGED INDEXED DECAYED PROPAGATED 01 · TEST Test Variants, audiences, bids — live across every channel 02 · TRACK Track ROAS, CAC, fatigue — every signal, every minute 03 · OPTIMIZE Optimize Budgets shift. Winners scale. Losers pause. 04 · SCALE Scale Winners spread — new geos, audiences, next loop
Watch the loop run on your data. Book a demo

Hypothesis creation at machine speed. Testing, learning, and allocating every dollar against the KPI you actually report on.

Responsibilities · Daily / Weekly 22 active
  • Reallocate budget across campaigns $24KMOVED · 24H
  • Test 10+ hypotheses per week 22THIS WK
  • Mine insights from performance patterns 14NEW · WK
  • Bids, pacing, frequency — in real time 15MCADENCE
  • Creative rotation against fatigue index 7SWAPS · WK
ROAS · 7-week journey 4.1× ▲
START 1.2× NOW 4.1× 7-WEEK LIFT +2.9× 4.0× 3.0× 2.0× 1.0× W1 W2 W3 W4 W5 W6 NOW WHAT MOVED THE LINE LAL +3% Freq cap 3 × Age 45+ CPA +20% Hero swap Bid AM +15% WON × KILLED
22
Hypotheses
14
Positive
310
Decisions · 24h
Live stream Hypotheses arriving · accepted or rejected in real time
Off-peak bid +15% on top-3 creatives n=2,184·conf 0.71
Testing
Frequency cap = 3 · broad audiences CAC −24%·conf 0.93
Accepted
Age 45+ broad · CAC $48 over bar n=4,820·save $1.2k/d
Rejected
Lookalike +5% expansion · D2C IN n=1,612·conf 0.64
Testing
Hero swap v2 → v3 · Reels 6s CTR +38%·conf 0.91
Accepted
Reels weight ×1.6 over Feed CPM −38%·+$2.4k/d
Accepted
Trend audio swap · baseline CTR n=486·no lift
Rejected
Geo-weight Tier-1 cities · cap $28 n=974·conf 0.58
Testing
CTA shorten · 9 → 4 words CVR +18%·conf 0.86
Accepted
"First-time only" hook · CVR +2.4× scaled 1 → 4 sets·conf 0.94
Accepted
Daypart cap weekends · flat ROAS n=2,108·no signal
Rejected
Music tempo +20 BPM Hold 2.1× longer·conf 0.88
Accepted
Off-peak bid +15% on top-3 creatives n=2,184·conf 0.71
Testing
Frequency cap = 3 · broad audiences CAC −24%·conf 0.93
Accepted
Age 45+ broad · CAC $48 over bar n=4,820·save $1.2k/d
Rejected
Lookalike +5% expansion · D2C IN n=1,612·conf 0.64
Testing
Hero swap v2 → v3 · Reels 6s CTR +38%·conf 0.91
Accepted
Reels weight ×1.6 over Feed CPM −38%·+$2.4k/d
Accepted
Trend audio swap · baseline CTR n=486·no lift
Rejected
Geo-weight Tier-1 cities · cap $28 n=974·conf 0.58
Testing
CTA shorten · 9 → 4 words CVR +18%·conf 0.86
Accepted
"First-time only" hook · CVR +2.4× scaled 1 → 4 sets·conf 0.94
Accepted
Daypart cap weekends · flat ROAS n=2,108·no signal
Rejected
Music tempo +20 BPM Hold 2.1× longer·conf 0.88
Accepted
14 Accepted today
9 Rejected today
9 Currently testing
FeatherDB · Living Context

The memory underneath the agent. What your account learned last quarter, working for you today.

Feather is where the agent stores everything it learns about your account: winning hooks, audience patterns, what worked when. Every test feeds the next. Every dollar spent makes the next dollar smarter. Same mistake, never twice.

Feather · account.feather
1.2 GB · embedded
~/account.feather 5 directories · living
brand/last write 14s
voice.toneweight 0.94
guardrails.json48 rules
creative.dna9 patterns
audience/decay τ=14d
segments.idxlive
recency.weightsre-weighted
cohorts/37 active
creative/multimodal
winners/148 indexed
fatigue.idx0.62
multimodal.bundletxt + vis + audio
spend/outcomes · attribution
memory/600+ patterns · queryable
01Capture

Capture

Every winning hook, audience, bid, and outcome gets tagged and stored. Hooks, openers, music, CTAs, language, geo, audience, ROAS, CPA.

02Running Context

Running Context

The live state of what's working right now. Queryable by the agent in real time, across markets and channels. Not a dashboard. A substrate.

03Decay & Stickiness

Decay & Stickiness

Hot patterns stay hot. Stale ones decay. The system distinguishes "worked once" from "works repeatedly," and acts accordingly.

04Propagation

Propagation

Winning patterns from one geo seed tests in another. Knowledge moves laterally, without a human carrying it from spreadsheet to spreadsheet.

FeatherDB · Living Context

The memory underneath the agent. Here's what it brings to your account.

Memory

Every winning creative, every decision, every outcome stays remembered across launches. Never rebuilt from scratch.

Learning

Every test feeds the next. Same mistake, never twice. The agent gets sharper each week, not reset.

Compounding

Every dollar spent makes the next dollar smarter. Insights compound across channels, accounts, seasons.

Context

Brand, audiences, creatives, history: queryable, always. Every new campaign starts with the full picture.

Step 01 Intelligence Signals
1,000+ · captured / week
Creative Hook Patterns

Top-converting hooks across categories, formats, languages.

Continuous
Competitor Intel

Meta & Google ad libraries: creatives, hooks, spend signals.

Daily
Trends & Virality

Sounds, formats, viral mechanics across platforms.

Hourly
Audience Behavior

Install, retention, paid-conversion by cohort & geo.

Real-time
Brand Corpus

Voice, tonality, vertical guardrails per category.

Weekly
Performance History

18 months of CPA / ROAS / LTV by audience & geo.

Always-on
Plug Feather into your account. Book a demo

Agents do the labor. You do the judgement.

Hawky doesn't replace your team. It changes their job. The agent surfaces actions, you accept, reject, modify, or command back. Every decision passes through your rails.

Agent suggestions 5 awaiting
Live · auto-streaming
#AGT-2841 Tactical Pending
Raise AM bid +15% on top-3 creatives.
Trigger CPM −8% · conf 0.91
Impact +$2.4k proj. · 1-click reversible
Your decision · click any action
0 Acted
0 Rejected
0 Modified
0 Learned
Recent decisions
  • Click an action to start. Decisions show up here. Idle
Run by the agents

Automated · 24/7

  • A1Variant generation across image, video, and copy at scale
  • A2A/B test execution & rotation across audiences
  • A3Fatigue detection per channel, geo, audience
  • A4Spend reallocation across campaigns & channels
  • A5Hypothesis testing across bids, frequency, and audiences
  • A6Competitor scrape, response, & insight surfacing
  • A7Performance reporting into Feather
Gated by your directors

Human · Judgement

  • H1Brand narrative & category positioning
  • H2Brand voice rules & sensitive-category vetoes
  • H3Channel mix & budget caps per market
  • H4Escalation thresholds for anomalies and regulatory flags
  • H5Quarterly creative strategy & vertical priorities
  • H6Final brand-pass on launch & flagship creatives
  • H7Approval of new audiences, geos, partnerships
Take the agent for a spin. Book a demo

We only get paid when the agent delivers.

Performance-based pricing.
Tied to your results, not our presence.

Talk to us about what success looks like for your accounts. We'll build the commercial around it.

Run it yourself, or let our team run it for you.

You run the agent

Your team stays in control. Plug into your accounts, set guardrails, watch the agent work. Best for in-house teams that want speed without giving up the wheel.

  • Direct integration with your ad accounts
  • Configurable guardrails & spend caps
  • Decision log + reversible moves
  • Slack & email alerts

We run it for you

Results from week one, zero learning curve. Our performance marketers operate the agent on your behalf: strategy, guardrails, reporting, all included. Best for teams that want outcomes, not operations.

  • Dedicated growth pod
  • Weekly strategy reviews
  • Monthly outcome reports
  • SLA on response & reporting

Everything buyers ask before they switch.

Detailed answers for performance marketers, CMOs, RevOps, and the engineering leads they bring in for procurement. Click any question to expand.

The basics · What is a Performance Marketing Agent?
What is a Performance Marketing Agent?

A Performance Marketing Agent is an AI agent that runs paid media accounts end-to-end. It launches campaigns, monitors performance, tests hypotheses, optimizes budgets and bids, and scales winners. Same judgement loop a senior media buyer uses, at machine speed and 24/7. Hawky's Performance Agent is purpose-built for this. It doesn't just analyze data or generate creative, it operates the account in-platform, with every move logged and reversible.

How is a Performance Agent different from "AI for ads" tools?

Most tools surface insight (dashboards) or generate one thing (a creative variant, a recommendation). The Performance Agent does the job. It executes a full closed loop: Test → Track → Optimize → Scale. It replaces the work, not just the work's reporting layer. There's another piece too: Hawky stands on Feather's Living Context, so it gets sharper week over week instead of resetting between campaigns.

Is this software (SaaS) or is it actually an agent?

It's an agent. The distinction matters: software charges you for access to features. An agent charges you for outcomes. Hawky isn't a dashboard you log into. It's a teammate that runs the account every minute the auction is open. You see what it does, you can override it, you can roll it back. But the work happens whether you're at your desk or not.

What does the agent actually do every day?

On a typical day across one account, the Performance Agent will: monitor 2,000+ events per hour across Meta/Google/YouTube; test 2–4 new hypotheses (audiences, bids, creatives); kill or scale prior tests based on confidence thresholds; rebalance budgets across campaigns; rotate fatigued creatives; surface 1–3 new patterns from competitor & trend signals; log every decision with the data that triggered it.

Who is the Performance Agent for?

Heads of growth, CMOs, performance marketing leads, and growth engineers running $50k–$5M+ monthly ad spend across Meta, Google, and YouTube. Most useful for D2C, fintech, EdTech, quick commerce, mobility, and any high-velocity performance category where creative testing is the primary lever.

Trust & safety · Is it trustable?
Is it safe to give an agent write access to my Meta & Google ad accounts?

Yes. The safety model is the same one your in-house team operates under, made stricter by guardrails. The agent connects via Meta and Google's official APIs (the same OAuth flow you use for any reporting tool). You set spend caps, daily-change ceilings, allow-listed campaigns, and escalation thresholds. The agent cannot touch anything outside the rails you draw. Every action is logged with the trigger and is one-click reversible. SOC 2 Type II.

What if the agent makes a bad call?

Three protections, in order. Guardrails: the agent can't take actions outside the rules you set (max bid changes per day, no single creative pause without 95%+ confidence, etc.). Logging: every action carries the trigger, the supporting signal, and a confidence score. Reversibility: any move can be rolled back in one click, individually or as a batch. Most customers configure auto-pause if any 24h window underperforms a baseline by >X%.

How does the agent handle brand safety?

Brand voice rules, sensitive-category vetoes, and creative blacklists are configured by your brand lead and live in Feather as guardrails the agent must respect. Final brand-pass on launch and flagship creatives is gated by a human (see Human-in-the-Loop). The agent will not generate, push, or scale creative outside the brand corpus you've defined.

Where does our data live? Is it used to train models?

Your account data lives inside Feather, in your environment. .feather is one embedded file. No third party in the loop, no data leaving for training. We do not train foundation models on customer data. Aggregated, anonymised pattern signals (e.g. "off-peak bid lifts work in IN D2C") may flow to the broader pattern library only with explicit opt-in.

Can humans override the agent in real time?

Always. Every action shows up in the Decision Log within seconds. You can pause the agent globally, pause it per campaign, or roll back any single move. Many customers run a "shadow mode" for the first 2 weeks where the agent suggests every move but a human confirms. Once trust is built, they switch to autonomous on a per-channel or per-campaign basis.

Is this trustable for big spend ($1M+/month)?

Yes. That's the sweet spot. The bigger the spend, the more leakage from manual ops, and the more the agent's 24/7 coverage compounds. Customers in the $1M to $5M/month band typically see the largest absolute ROAS lifts, because percentages compound on bigger denominators. Spend caps and per-channel caps make this safe.

Human vs agent · What stays with my team?
Will this replace my performance team?

No. It changes their job. The manual lifecycle (variant generation, A/B execution, fatigue detection, spend reallocation, hypothesis testing, competitor scraping, reporting) moves to the agent. Your creative directors, brand leads, and performance leads stay in the seats that matter most: narrative direction, brand voice rules, channel mix, escalation thresholds, quarterly strategy, final brand-pass on flagship creatives, and approval of new audiences, geos, or partnerships.

What's the boundary between agent work and human work?

See the Human-in-the-Loop section above. As a rule of thumb: agents do the labor; humans do the judgement. Anything tactical, operational, repeatable, and bounded by rules → agent. Anything strategic, brand-facing, or involving new categories of risk → human.

Do I still need a media buyer or agency?

You still need strategy, brand, and oversight. You need fewer hours of routine ops. Most customers retain a senior performance lead who acts as the agent's manager: setting strategy, reviewing weekly, approving new audiences. Junior buying roles often shift toward strategy and creative oversight rather than dashboards. Agencies that move toward strategic and creative work continue to add value. Pure execution agencies are the ones replaced.

Build vs buy · Why not just use Claude / GPT directly?
Why not just build this with Claude or GPT ourselves?

You can, and a few teams do. The build cost is rarely the LLM API. It's the ad-platform integrations (Meta, Google, YouTube each have non-trivial quirks), the safety layer (guardrails, escalation, audit logs, rollback), the memory layer (patterns that decay, stickiness, propagation across geos: Feather), the evaluation harness, and the on-call. Most teams that try this discover they've taken on 6 to 9 months of engineering work that doesn't differentiate their product, just to reach feature parity. Hawky exists so you don't have to.

Can we self-host or run Hawky inside our infrastructure?

Yes, for enterprise customers. Feather is embedded by design (one file), so the substrate runs in your environment with zero outbound. The agent itself can run in your cloud or as a managed deployment. Talk to sales for the deployment options that fit your security review.

How is this different from Pixis, Smartly, AdCreative, Pencil, Omneky?

The short version: most of those tools generate creative variants or surface optimization recommendations. They don't do the work. They hand a recommendation back to a human. Hawky is end-to-end: it generates, tests, runs, optimizes, scales, and logs. We have a longer competitive teardown. Happy to share on a call.

Do I need engineers to set this up?

No. Connection is OAuth-based, so your performance lead can do it in 15 minutes for Meta, Google, and YouTube. If you want custom guardrails or escalation hooks (Slack, email, custom webhook), you'll want an engineer for an afternoon. Fully Managed customers don't touch setup at all.

Results & timeline · How fast do we see it work?
How fast do we see results?

Managed customers typically see directional signal in week one (decisions logged, fatigue caught, overnight wins captured). DIY customers see meaningful KPI shifts inside 30 days. Most customers see ROAS uplift inside 90 days. There's no warm-up tax. The agent learns from your historical context the moment Feather is connected.

What kind of ROAS uplift should we expect?

Cohort median across 200+ active customers: +25% ROAS in the first 90 days, −43% CAC on competitive categories, +10× hypothesis testing velocity. Range is wide. Accounts with poor manual hygiene see larger lifts. Well-optimized accounts see smaller percentage lifts but larger absolute gains. We measure against your own pre-Hawky baseline, not industry benchmarks.

How does the agent learn over time?

Through Feather's Living Context. Every test outcome, every winning creative, every audience that worked or didn't gets captured, decayed by recency, and propagated to similar accounts or geos when a pattern is statistically meaningful. Week one the agent is good. Week twelve the agent has 600+ patterns specific to your brand and is dramatically better than week one. It compounds.

What happens if I cancel?

You keep your data, your account state, your decision log, and a Feather export. The agent disconnects from your ad accounts. No lock-in, no data hostage. We'd rather you stay because it's working than because leaving is hard.

Pricing & engagement
How does pricing work?

Performance-based. The commercial is built around your KPI, not our headcount. We share specifics on a 30-minute call after we know your spend, channels, and KPI definition. No flat retainers. No per-seat fees.

Can we run a pilot?

Yes. Most enterprise engagements start with a 30-day pilot on one account or one channel. We agree the success metric upfront, run the pilot, and the pilot fee converts into the main engagement if you continue. If we miss the metric, the pilot is on us.

What's the difference between DIY and Fully Managed?

DIY: your team operates the agent. Best for in-house performance teams that want speed without giving up the wheel. Fully Managed: our growth pod operates the agent on your behalf. Strategy, guardrails, weekly reviews, monthly outcome reports. Best for teams that want outcomes, not operations. You can switch between modes without re-onboarding.

How long does onboarding take?

15 minutes to connect the ad accounts. 24 hours for Feather to ingest 18 months of historical context. 7 days for the agent to ramp into autonomous mode (or longer in shadow mode if you prefer). Most accounts have meaningful agent-driven wins logged inside the first week.

Feather & architecture
What is Feather, and how is it different from a CDP or warehouse?

Feather is the Living Context layer the agent stands on. CDPs are static (they snapshot the world). Warehouses don't decay (yesterday's truth weighs the same as today's). Vector databases don't know your brand. Feather is a living substrate. Patterns are tagged, indexed, decayed by recency, and propagated where they matter next. It's MCP-native, embedded as one file, and your data never leaves.

Is Feather a separate product?

Yes. Feather is its own product. Hawky is its first consumer. Other agents (and your own internal agents) can plug into Feather as the substrate. If you want to talk to us about Feather independently, we're happy to.

What channels does it run on today?

Meta, Google, YouTube on day one. The rest of the perf-marketing stack (TikTok, LinkedIn, Reddit, Apple Search Ads, programmatic) is on the roadmap. Living Context preserves account memory across all of them as they come online, so there's no re-learning when a new channel is added.

How do you handle attribution?

The agent reads attribution from the platforms you already use (Meta CAPI, GA4, your MMM if you have one). It does not make attribution decisions for you. It operates against the KPI you report on, with the attribution model you trust. We can also plug into your own warehouse for blended ROAS calculations if you have one.

Feather deep dive · Living Context for performance marketing
What is Feather, exactly?

Feather is the Living Context layer that AI agents plug into for performance marketing. It's a living substrate, meaning the data has decay, stickiness, and multimodality built in. Shipped as one embedded file, MCP-native, where your data never leaves your environment. Hawky's Performance Agent reads from and writes to Feather. Other agents (yours included) can plug in the same way.

How is Feather different from a regular database?

A traditional database stores rows. A vector DB stores embeddings. A warehouse stores history. None of them decay the way performance signal does. Yesterday's CPM win matters less than today's, and last quarter's pattern only matters if it keeps mattering. Feather treats memory as living: stickiness for things that keep working, decay for things that don't, multimodality (text, visual, audio) under one entity ID. It's not a better Postgres or pgvector. It's a different category.

Is Feather a vector database, CDP, or data warehouse?

None of those, and that's the point. Vector DBs are commodity infrastructure. Feather operates above them. CDPs are static. Feather is living. Warehouses don't decay. Feather does. If you slot Feather into one of those buckets, you'll mis-evaluate it. The closer mental model is "Performance Context Layer": purpose-built substrate for the way agents actually work.

Can I use Feather with Claude, GPT, or my own agents?

Yes. Feather is MCP-native (Model Context Protocol, Anthropic's open standard). Any agent that speaks MCP can read and write Feather as a tool, not a dataset. That includes Claude, your own Claude Agent SDK builds, GPT via MCP shims, and internal tooling. Hawky is the first consumer. Other agents are the next.

What does "MCP-native" mean for me?

It means Feather presents itself to your agents as a set of tools, like `query_audience`, `recall_winning_creative`, `get_brand_voice`. Each one has typed inputs and outputs, the same way the agent calls the Meta API or sends a Slack message. You don't write SQL. You don't manage embeddings. The agent treats Feather like any other tool in its toolbelt, and Feather handles decay, indexing, and propagation underneath.

Why is Feather "embedded" / one file?

Feather ships as a .feather file that lives inside your environment. No SaaS round-trip. No "send your performance data to a third party for inference." This matters for two reasons: sovereignty (your data never leaves, nothing trains on it, auditing is trivial) and speed (queries don't pay network latency, so the agent gets sub-millisecond context lookups).

Can I use Feather without Hawky?

Yes. Feather is its own product. Hawky is the first consumer. You can be the next. Teams plug Feather into their internal agents, in-house performance tooling, or even Claude-based workflows. If you'd like to talk to us about Feather independently of the Performance Agent, we're happy to.

What's actually in Feather for a performance marketing account?

Six categories of living context: Creative Hook Patterns (continuous), Competitor Intel from Meta and Google ad libraries (daily), Trends & Virality across platforms (hourly), Audience Behavior by cohort (real-time), Brand Corpus with voice and category guardrails (weekly), and Performance History covering 18 months of CPA / ROAS / LTV by audience and geo (always-on). Each lives through the four-stage lifecycle: capture → running context → decay & stickiness → propagation.

Does Feather work with my existing data warehouse / CDP?

It complements them. Your warehouse remains the system of record for raw events. Feather is the system of working memory the agent uses to make decisions. We can ingest from Snowflake, BigQuery, Redshift, Databricks, or any reverse-ETL pipeline. Feather doesn't replace your CDP. It sits above it, decaying what doesn't matter and surfacing what does.

What's the difference between Feather and Hawky?

Hawky is the agent: the operator that runs your accounts. Feather is the substrate: the memory layer the agent stands on. Hawky pays for Feather. Feather makes Hawky compound. They're a matched pair, but Feather is its own product with its own audience (AI engineers building performance tooling). Don't collapse them. Buying one doesn't lock you out of the other.

Get Started

Hire the agent.
Pay on results.

30-minute demo. See the loop running on a sample account before you sign anything. Or jump straight to the free 30-day pilot below.

30-day pilot · free
Make your marketing org AI-native.

Run the agent on one account for 30 days. If it works, scale across every account. Your team trades manual ops for agent operators.

  • 15-min onboarding. OAuth · no engineering.
  • Real loop on real data. No sandbox.
  • Outcome-priced. Pay on your KPI.
  • Reversible. Every move, rollback-able.