In his first month, DeepReachAI founder Aryan Agrawal focused on Reddit and X, reading complaints about cold outreach and replying one by one rather than adding product features. In an IndieHackers post, he said every signup came from there, none from a formal launch post, and ad spending was zero. The title read $150 MRR, 250 users, and $0 spent on ads—all his own figures at the time. [1]
DeepReachAI handles cold outreach: find companies and people, research recent developments, then generate email, LinkedIn, or WhatsApp messages. Aryan originally thought the personalization engine would be the hook. His account in the post was the opposite: what persuaded people to connect accounts was “we don’t auto-send LinkedIn, we don’t scrape WhatsApp.” In his words, compliance mattered more than the AI part.
That sentence explains the product’s current form better than any feature list on its website—but only half of it. There are several layers to where it stops, and one has not yet aligned.
It stops in more than one place
Orio is the product’s execution workflow; the middle of the website places it in the Enterprise tier. It starts only after an explicit outreach instruction. It searches for hiring, project, contract, and request signals, sending only qualifying companies into contact enrichment. Names, email addresses, and phone numbers appear if they can be resolved. Otherwise, they remain blank, without filling in or guessing. [2]
There are two sets of rules for sending. With the standard generator and CSV sequences, drafts open in Gmail compose. The person reviews and presses Send, then returns to History to mark the message sent before the follow-up timer starts. Orio differs: under the website’s Step 4, an explicit outreach instruction leads it to save LinkedIn drafts and, where an address is available, send the researched first email through a connected account. LinkedIn remains draft-only; it does not send for you. The website FAQ is more direct: the standard generator and CSV sequences do not send every generated message, but prepare drafts for review.
According to the homepage, “a person presses Send” is the standard workflow’s rule; Orio is the exception authorized to send a first email. These remain feature claims, not verified usage.
Image: Browser capture from the official page[3] on 2026-11-18, retaining the original interface. The example is not a hands-on test from this research.
When email replies arrive, it classifies them into four categories: WhatsApp opt-in, call consent, rejection, or ambiguity. An ambiguous reply receives one follow-up after three days—only one. A rejection stops that path. WhatsApp waits for opt-in. A call requires all four conditions: explicit consent, an available number, an active calling connection, and remaining plan minutes. Once a time is agreed, it checks the connected calendar, creates an event, and invites the contact.
These “no” rules are boundaries expressly written into the product workflow. According to Aryan’s post, they are precisely what made people willing to connect accounts; he credited compliance rather than AI. As for what the limits have prevented for real customers, the public material supplies neither customer outcome cases nor verifiable results data.
A builder still deciding how far to go
On the About page, he describes himself as a data scientist and AI automation engineer in Pune. Through AutomateIQ, he also builds automation systems for other companies, personally operating with a small two-person team, with the builders fixing problems themselves. [4]
His IndieHackers post described a pre-seed, solo operation with no team yet, and said he was still weighing how far to push automation versus keeping it lean. The two descriptions differ, and the post displays no date to establish their order. They may describe changing stages, rather than a simultaneous contradiction.
The uncertainty in that post about how far to go shows only that he was weighing automation boundaries at the time. The website now presents the story as “from a cold name to a booked meeting, fully handled,” and About describes a full-outreach AI Employee as the long-term goal. At the same time, calls and WhatsApp retain consent gates, and LinkedIn remains draft-only. The end-to-end marketing promise and those stopping conditions coexist.
Who pays, and what is metered?
The middle of the website and its lower summary do not agree on which plans include Orio, Find Leads, contact enrichment, and sequences. What can be established is that these actions consume credits: Free resets weekly, Starter and Pro daily, and Enterprise’s Starter and Pro weekly. Unused allowances expire at reset without rollover. The terms price additional credits at ₹1 each in India or $0.01 elsewhere, with a minimum purchase of ₹20 or $1. Purchased credits are used after plan allowances. If a platform-side fault makes an action fail, its credits are automatically returned. When credits run out, Orio campaigns pause and resume after a top-up. According to the website FAQ[2], periodically issued plan credits are nonrefundable; refunds for purchased extra credits require contacting support for manual review. The terms say subscription-fee refunds require contacting the company, which decides whether there is a valid reason; paid plans do not automatically refund merely because quotas went unused. Purchased extra credits do not expire, unlike the periodically expiring plan allowance. [5]
Neither the homepage nor the terms list subscription prices for the tiers; they point to a Pricing page. That entry requires login, so logged-out visitors cannot see specific prices. Subscription prices have not been publicly verified.
For the same category of work, RegieGO also sends email from users’ own Gmail or Outlook accounts. Free is $0, with 1,000 one-time credits per person pooled across up to ten people. Pro costs a $41/month equivalent with annual billing or $49 billed monthly, including 5,000 credits a month. Contact enrichment is charged only when an email address or phone number is actually found. [6] · Waterfall Lead Enrichment[7] For calls, RegieGO uses its own parallel dialer with local presence, DNC, and up to 5 lines. About · Regie.ai[8] DeepReachAI places calls behind explicit consent, an available number, an active calling connection, and plan minutes. Both address when a call may be made, but put the controls in different places. The public material provides no outcome comparison.
One point that does not add up
The homepage says Orio sends a researched first email through a connected account when an address is available. Yet §11 of the terms describes Gmail sending as an optional feature “if released,” while the privacy policy says the current repository does not implement Gmail OAuth. Privacy policy[9] The homepage does not say whether the “connected account” is Gmail. If it is, the different authorization states need explanation. If another channel is used, the public material does not identify it. The public pages leave that scope question unanswered.
Gaps between pages are hardly news for a product reporting a little over two hundred users. But this gap lands on the core promise: who presses Send? In Aryan’s account, users connect accounts because they trust the boundary of “no automatic LinkedIn sending, no WhatsApp scraping.” If authorization for email sending remains expressed only as a vague “connected account,” the boundary being sold is not fully grounded yet.
As for money: $150 MRR, 250 users, and $0 ad spending are the author’s historical self-reported figures in that post, not current revenue and certainly not profit. Cost structure, acquisition cost, and customer distribution remain unquantified. In the post, he put the typical industry cold-outreach reply rate in the 3–5% range, with most tools stopping at lead finding or email sending.
It is that boundary that makes people willing to connect accounts. The first test of the boundary is whether the homepage, terms, and privacy policy agree.