Skip to content

Free for one month, then $2 per seat per month.

That is the pilot, and we’re choosing five companies for it; three places are open. The $2 rate is locked in for a year, and you can leave any time. After the pilots, Billy moves to standard pricing, which we will publish and tell you about before it applies to you.

Leave any time

  1. FreeThe first month

  2. $2 per seat per monthLocked in for a year

  3. After the pilotsStandard pricing

Billy is taking on five pilot companies while its security reviews are in progress. That is why, for now, it starts with a short call. Security reviews.

Seats and usage.

A seat is one person who is allowed to use Billy. If there are people you do not want Billy to respond to, we do not count them as seats.

In the pilot, each seat adds 2.5 million tokens a month to one pool that the whole company shares. The pool counts the input and output tokens of the model calls Billy makes. If you need more, $20 adds 25 million tokens to the pool, for the whole company and not per seat.

What we sell is a seat. Tokens are a cost to us, not our product, so Billy is built to use few: a recurring job is planned once, then runs as software, and calls the model only where a run needs it. In the 30 days to 17 November 2026, the runs of all our own team’s recurring jobs together made 71 model calls and used 0.16 million tokens.

The admin page shows usage task by task: the tokens each task used, its model cost, and which person it was for.

Screenshot of an admin page, with six green boxes drawn over its table, labelled “Each task”, “Where”, “Who”, “What it did”, “Tokens” and “Cost”. Above the table stand the heading “Sessions”, the line “click a row to open the full call timeline”, a search field, a field “From: anywhere”, a filter set to “Arjun Jain”, a button “Errors only” and the menus “Any cost” and “Sort: recent”. The columns are Session, From, User, Started, Duration, Calls, Err, Skills, Tokens and Cost; Cost is the model cost of the session, not a charge. Six rows; in each the user is Arjun Jain and Err shows a dash. Session 22882204: #team-billy-real, 11-17 09:47, 1h 10m, 11 calls, skills “intent-classifier” and “planner.operations”, 234.2k tokens, $0.0244. Session 22884540: #team-billy-real, 11-17 09:50, 29s, 7 calls, the same two skills, 237.3k tokens, $0.0132. Session 22875249: #team-video-production, 11-17 08:40, 0s, 2 calls, skill “intent-classifier”, 13.7k tokens, $0.002106. Session 22789414: DM · Arjun Jain, 11-16 21:43, 28s, 6 calls, the same two skills, 190.4k tokens, $0.0141. Session 22788954: DM · Arjun Jain, 11-16 21:41, 46s, 8 calls, the same two skills, 830.8k tokens, $0.0431. Session 22788956: #site-launch, 11-16 21:41, 23s, 8 calls, the same two skills, 270.7k tokens, $0.0157.
FastCode AI’s own admin page on 17 November 2026: the sessions table, filtered with the page’s own User filter to one person, Arjun Jain, our founder. Six of Arjun’s most recent sessions, of 16 and 17 November, asked in three channels and in Arjun’s DMs with Billy, each with its start time, duration, model calls, skills, tokens and cost: the model cost of the task, not a charge to you. These token totals include cached reads, which the allowance does not count: the 46-second session shows 830.8k tokens, of which 144.7k are input and output, the part the allowance counts. Nothing is blurred. The green boxes, their labels and the lines that join them are ours, drawn over the picture; no number or text of the page was changed. Left out: the rest of the page.

What is included.

The pilot is company wide. It includes:
  • One question searches every source at once
  • Takes the notes of each call: decisions, action items and their owners
  • Set up recurring jobs by asking in plain English
  • Keeps Slack Lists up to date: owners, status and due dates
  • Can keep a link on a list row to the message where it happened
  • Runs polls and chases teammates for replies
  • Billy reads your public channels and the private channels it’s added to, and won’t tell you what was said in a channel you aren’t in
  • Can show you a card to approve before it acts
  • Logs every action in a hash-chained log
  • Billy works in English, Japanese, Korean and German

You work directly with the people who build it.

What others charge.

Billy’s price is published. Three of the ten tools we compare don’t publish one.

See the full comparison
  • Billy

    Pilot, for five companies we choose: free for one month, then $2 per seat per month, locked in for a year. Each seat adds 2.5 million tokens a month to one pool the whole company shares, and $20 adds 25 million more. Leave any time.

  • Instinct

    Not published (read on 17 November 2026)

  • Muse

    Free for most of what people need, with subscription plans; its home and business pages name no price (read on 17 November 2026)

  • Claude Tag

    Included in Anthropic’s Team and Enterprise plans. Team: $20 per seat per month if billed annually, $25 if billed monthly (Standard seat). Enterprise: $20 per seat per month, billed annually, 20-seat minimum, plus usage at API rates. Claude Tag itself adds no per-seat charge: its work in channels is billed by usage, from a balance the company funds, and people need no seat to use it there. No per-task cost is published (headline and price read on 17 November 2026)

  • Glean

    Not published; book a demo

  • Lindy

    From $29.99 per user per month with 3,000 credits; extra credits $10 per 1,000 (read on 16 November 2026)

  • Sentra

    Team, for 2 to 20 people: $16 per user per month, free as a launch special, with 2,000 action credits per user per month; Enterprise custom (read on 16 November 2026)

  • Hyperspell

    Not published; book a demo

  • Viktor

    Credits; free start with up to $100 in credits; Team $100 a month with 40,000 shared credits (read on 16 November 2026)

  • Artie

    Free with up to 14,000 credits; Team $100 per month, usage-based, no per-seat fees (headline and price read on 16 November 2026)

  • Slack AI

    Thread, channel and huddle summaries from Pro; Slackbot on Free and Pro as “a limited trial experience” (Slack’s FAQ) and consumption-based, 15 messages per user a week on Business+ and unlimited on Enterprise+; AI search from Business+ and enterprise search on Enterprise+. No separate AI price; higher limits through sales (read on 16 November 2026)

Prices and headlines were read from each company’s own website on 7 November 2026, unless an entry gives a later date. Product names belong to their owners.

Testimonials

What people say about Billy

Abhijeet Katte (opens in a new tab)

Founder, CMUL8 (opens in a new tab)

CMUL8 is another company and our second pilot company. Its founders are friends of ours.

My attention jumps between a lot of things, and remembering every little commitment is hard. Billy keeps track of the follow-ups and reminds me when they matter, so I don’t have to carry everything in my head.

Up and running within a day.

Book a call with us. We’ll show you Billy live.

Book a call (opens in a new tab)