~ / AI Research / Building SaaS on Bluesky & AT Protocol
Can you build a real SaaS business on top of Bluesky and the AT Protocol in 2026? This analysis covers the platform’s current state (40M+ users, ~$0 revenue, 25 employees), the technical foundation that makes it uniquely developer-friendly (free API, open firehose, algorithmic choice), the existing ecosystem of 50K+ custom feeds and dozens of third-party tools, the historical precedent from Twitter’s early ecosystem (Buffer $22.6M ARR, Sprout Social $455M revenue, TweetDeck $40M acquisition—then the API apocalypse that killed Tweetbot, Apollo, and hundreds of others), what SaaS products could work, the real risks (engagement down 50% from peak, 18% 30-day retention, Threads at 450M MAU), and the bootstrapper playbook for building on an open protocol.
The core question: The AT Protocol is structurally different from Twitter’s API—it’s an open protocol, not a platform permission. But is 40M users (with ~4M DAU) enough to sustain a SaaS business? And can you build before Bluesky builds it themselves?
| Metric | Number | Notes |
|---|---|---|
| Total registered users | ~40.2 million | Reached Nov 2025; likely 41–43M by Feb 2026 |
| Daily active users (DAU) | ~3.5–4.1 million | ~8–11% of registered users show up daily |
| Monthly active users (MAU) | ~27 million | Based on Jan 2025 global MAU data |
| Daily unique posters | ~1.3 million | Active content contributors |
| Total posts ever | 658 million+ | As of late 2025 |
| Lifetime app downloads | 26.2 million | iOS + Android combined, as of Nov 2025 |
| Date | Users | Key Event |
|---|---|---|
| Feb 2023 | Invite-only launch | Jack Dorsey-backed project opens waitlist |
| Feb 2024 | Public launch | Open to all, no invite needed |
| Aug 2024 | ~10 million | Brazil’s X ban drives 420K downloads/day |
| Sep 2024 | ~10 million | Steady state |
| Oct 2024 | ~13 million | Musk weakens X’s block feature; +500K in one day |
| Nov 6, 2024 | ~15 million | US election: largest X exodus since Musk takeover |
| Nov 25, 2024 | ~22.5 million | +7.5M in 19 days (peak growth) |
| Mar 2025 | ~30 million | Post-election settling |
| Aug 2025 | ~38 million | Growth slows to ~1.6M/month |
| Nov 2025 | ~40.2 million | 302% increase from Sep 2024 |
| Feb 2026 | ~41–43 million (est.) | Steady additions at ~0.5 users/sec |
Key pattern: Growth is event-driven, not organic. Each “X exodus wave” brings a spike, followed by significant churn. The platform went from adding ~5M users/month at peak (Nov 2024) to ~1.6M/month by mid-2025.
| Platform | MAU | DAU | Revenue | API Access |
|---|---|---|---|---|
| X (Twitter) | 540–570M | ~245M | ~$2.5B (declining) | $5K–$42K/month |
| Threads (Meta) | 450M | ~137–150M | Bundled with Meta ads | Limited API |
| Bluesky | ~27M | ~3.5–4.1M | ~$0 | Free, open firehose |
| Mastodon | <690K | <200K (est.) | ~$0 (nonprofit) | Free, per-instance |
Context: Threads doubled from 200M to 400M MAU in under a year. Mastodon peaked at 2.6M MAU in Nov 2022, now below 690K—a cautionary tale for decentralized platforms. X remains dominant by raw numbers despite post-Musk turbulence.
The Authenticated Transfer Protocol (“ATP” or “atproto”) is a decentralized social networking protocol developed by Bluesky Social PBC. Unlike traditional social platforms where a single company controls the data, algorithms, and access, AT Protocol separates these concerns into independent, interoperable services.
| Feature | AT Protocol (Bluesky) | ActivityPub (Mastodon) |
|---|---|---|
| Data model | Global firehose; aggregated content indexed centrally | Per-server inboxes/outboxes between accounts |
| Schema | Strongly-typed Lexicons with strict validation | Flexible ActivityStreams 2.0 / JSON-LD |
| Account portability | Full: signed data repos + DIDs; no old server cooperation needed | Limited: requires old server redirect; no data migration |
| Scalability | Aggregating Relays reduce load; designed for global scale | Server-to-server delivery; individual nodes can be flooded |
| Algorithms | User chooses from any custom feed algorithm | Chronological only |
| Global view | Full network firehose available to anyone | No global view; limited to connected instances |
| Developer experience | Predictable data formats; easier interoperability | Flexible but compatibility issues between implementations |
| Standardization | IETF Internet Draft published Sep 2025; working group charter Jan 2026 | W3C Recommendation since 2018 |
The AT Protocol consists of three main services:
Data flow: All PDSes → Relay → App View → App (your phone)
The firehose is an aggregated real-time stream of all public data updates in the entire network, accessible via
WebSocket at the com.atproto.sync.subscribeRepos endpoint. No authentication required.
Produces hundreds of events per second. Completely free to consume.
JetStream: A lighter-weight alternative introduced by Bluesky that relays all messages via WebSockets in JSON format for a fraction of the payload size. Official Jetstream instances available for free, no auth needed.
Who uses the firehose: Feed generators, labelers, bots, search engines, analytics tools, real-time monitoring tools (like Firesky.tv), generative art projects, trend analysis systems, and social listening tools.
did:plc — novel method developed by Blueskydid:web — W3C standard based on HTTPS/DNS| Feature | Bluesky / AT Protocol | X (Twitter) |
|---|---|---|
| Base cost | Free | $200/month (Basic) to $42,000/month (Enterprise) |
| Firehose access | Free, no auth required | $42,000/month for full archive search |
| Rate limits | 5,000 points/hour, 35,000 points/day | Varies by tier; severely restricted on lower plans |
| Read access | Most public data accessible without auth | Auth required for everything; read limits on all tiers |
| Third-party apps | Encouraged; official OAuth support | 100K token limits; third-party clients killed in 2023 |
| Philosophy | “Designed for third-party development” | “Pay for access or leave” |
| Language | SDK / Library | Notes |
|---|---|---|
| TypeScript/JS | @atproto/api (official), atcute | Works on Web, Node.js, React Native |
| Go | indigo (official) | BGS implementation; bot libraries available |
| Rust | ATrium, jetstream-oxide | Typed library for AT Protocol |
| Python | atproto (nano implementation) | Community maintained |
| Swift | ATProtoKit | For iOS/macOS development |
| Dart/Flutter | bluesky package | Supports almost all endpoints |
GitHub activity: The main bluesky-social/atproto repo has 8,900+ stars.
The organization has 37 repositories. Active developer community with discussions, grants, and conferences.
| Tool | What It Does | Pricing | Notes |
|---|---|---|---|
| ClearSky (clearsky.app) | Track followers, blocks, account health; DID lookup | Free | One of the most popular community-built tools |
| Atlas (bsky.jazco.dev) | Engagement-based social graph visualization | Free | Built by Jaz, community developer |
| Blueview (blueview.app) | Personal dashboard; post-level analytics (likes, reposts, replies, reach) | Free | Designed for individual creators |
| SkyStats | Aggregated performance metrics; follower trends; best posting times | Free | User-friendly for creators wanting high-level data |
| BlueSkyHunter | Analytics dashboard + scheduling + automated DMs | $15/mo (launch), normally $29/mo | 14-day free trial; launched Feb 2025 |
| BskyGrowth | Full growth suite: analytics, scheduling, feed optimization, engagement tracking | Paid (subscription) | Bluesky-specific growth platform |
| Bluesky Meter (blueskymeter.com) | Free real-time analytics for any Bluesky user | Free | Public stats viewer |
| Bsky Hub (bskyhub.com) | Free Bluesky analytics | Free | Community tool |
| Platform | Bluesky Support? | Revenue | Notes |
|---|---|---|---|
| Buffer | Yes (official partnership) | $22.6M ARR | Cross-posting, scheduling; Bluesky announced the partnership on its blog |
| CoSchedule | Yes | N/A | Full platform support including Bluesky and Mastodon |
| SocialBee | Yes | N/A (from $19/mo) | Create, publish, reshare; supports 10+ platforms |
| Fedica | Yes | N/A | Formerly Tweepsmap; social media marketing automation |
| PostBridge | Yes | N/A | Schedule + instant post; 9 platforms supported |
| Hootsuite | No | $350M revenue | Does not support Bluesky as of Feb 2026 |
| Sprout Social | Unclear / No | ~$455M revenue | No confirmed Bluesky integration |
| Tool | What It Does | Usage | Business Model |
|---|---|---|---|
| Graze (graze.social) | No-code custom feed builder + ad platform | 4,500 feeds, ~3,000 users; includes top feeds (News, Booksky) | Raised $1M (Betaworks, Salesforce Ventures). 200 feeds running ads at ~$1 CPM. 30% commission. First large-scale sponsored content on Bluesky |
| SkyFeed | Web client + custom feed builder (TweetDeck-like) | Multi-column dashboard; popular feed creation tool | Free |
| Bluesky Feed Creator (blueskyfeedcreator.com) | Create and publish feeds | 3,596 active published feeds (subset of platform total) | Free |
Graze is the standout story here. It is one of the only platforms monetizing custom Bluesky feeds, doing so with Bluesky’s blessing. The ads are not targeted by personal data but by feed topic. Feed operators choose which ads appear, how often, and at what price. This is the closest thing to an advertising ecosystem on Bluesky.
| App | Platform | Key Features |
|---|---|---|
| Graysky | iOS, Android | First cross-platform third-party client. View others’ likes, GIF support (Tenor), inline translation |
| Skeets | iOS, iPadOS | Accessibility-optimized; alternative interface |
| deck.blue | Web | TweetDeck-style multi-column interface. Home feed, notifications, custom feeds, lists. Multi-account + post scheduling |
| TOKIMEKI (tokimeki.blue) | Web, Android | Multi-column, multi-account. Extra features: bookmarks, mass threading, media viewers, feed customization |
| Skyblaze | Web | Enhanced web client |
| App | What It Is | Status | Notes |
|---|---|---|---|
| Skylight Social | TikTok-like short video app on AT Protocol | Launched (App Store + Google Play beta) | Backed by Mark Cuban. Co-founded by Tori White and Reed Harmeyer. Accounts portable across all AT Protocol apps |
| Streamplace | Livestreaming video service on AT Protocol | Active; livestreamed ATmosphere conference | Backed by Livepeer. Supports high-quality video, clips, uploads |
| PinkSky | Instagram alternative (photo sharing) on AT Protocol | iOS + Android | Built by Ramon Souza. Profiles, photo/video feed, Stories (24hr) |
| Flashes | Photo-focused Bluesky client | In development | Images over text |
| SkySpace | Customizable MySpace-style profiles on AT Protocol | In development | Edit profiles with music, interests, friends |
| WhiteWind | Long-form Markdown blogging on AT Protocol | Active | Free; stored in your PDS. Use Bluesky account to login |
| Frontpage | Decentralized link aggregator (Reddit-like) | Active | Upvoting, commenting, notifications. Built on ATProto |
| Spark | TikTok-like app | Available | Part of growing AT Protocol ecosystem |
| Blebbit | Discord + Reddit + Groups blend | Active | Community platform on AT Protocol |
embed.bsky.app generates embed code snippetsThis is the single biggest structural advantage. X charges $5,000–$42,000/month for API access. Bluesky’s entire API—including the full network firehose—is free and requires no authentication for read access. This eliminates the primary cost barrier that killed dozens of Twitter-based businesses in 2023.
Unlike Twitter, which progressively locked down its API from 2012 onward, the AT Protocol is architecturally designed to enable third-party apps. Custom feeds, labelers, clients, and entirely new social apps are first-class citizens of the protocol. Bluesky has funded developer grants, held conferences, and partnered with tools like Buffer.
66.5% male, 62% under 35, heavy representation of developers, academics, journalists, and tech workers. This is the exact demographic that early-adopts tools, gives feedback, and builds word-of-mouth. It mirrors early Twitter (2007–2010), which spawned a $100M+ ecosystem of tools before the API lockdown.
In February 2023, Twitter killed free API access, destroying Tweetbot, Twitterrific, and hundreds of other apps. Developers who built Twitter tools for years were suddenly without a platform. Many are now looking for alternatives. Bluesky is the most developer-friendly option.
If you build on Twitter’s API, Twitter can change the rules overnight (and has, repeatedly). The AT Protocol is an open standard, now being standardized through the IETF. Even if Bluesky the company disappeared, the protocol would continue. Your app talks to the protocol, not to a company’s proprietary API.
Early Twitter (2007–2012) spawned multiple successful businesses before the API crackdown:
Bluesky in 2026 feels like Twitter in 2008–2010: small but growing, developer-friendly, with a vibrant ecosystem forming. The critical difference is that AT Protocol is an open standard, not a proprietary API.
Most existing Bluesky tools are free, community-built side projects. Very few are monetizing. The ones that are (BlueSkyHunter at $15–$29/mo, Graze with $1 CPM ads) are early and underbuilt. There is massive white space for polished, paid tools.
| Company | Origin | Revenue / Exit | Lesson |
|---|---|---|---|
| Sprout Social | Social media management (Twitter-first) | ~$455M/yr revenue; public (Nasdaq: SPT); 13% YoY growth | Multi-platform diversification is key. Started on Twitter, expanded to all platforms |
| Hootsuite | Twitter management dashboard | $350M revenue; $690M valuation; $200M+ ARR | Early Twitter mover; expanded to become platform-agnostic. 18M customers |
| Buffer | Twitter scheduling tool | $22.6M ARR; first $2M/month in Nov 2025; $2.5M net income. 31% YoY growth | Bootstrapped, transparent (“Open” dashboard). Survived by diversifying beyond Twitter |
| Later | Instagram scheduling | $62.8M annual revenue; $276M+ raised; acquired Mavely for $250M | Vertical focus (visual content) before expanding |
| TweetDeck | Twitter power-user client | Acquired by Twitter for $40M (2011) | Platform acquisition risk: Twitter eventually hobbled TweetDeck in 2023 |
| Nuzzel | Twitter news aggregation | Acquired by Twitter (via Scroll) in 2021; shut down | Acqui-hire death. Features never fully integrated. A Nuzzel clone now exists for Bluesky |
| Company/App | What Happened | Revenue Lost | Lesson |
|---|---|---|---|
| Tweetbot (Tapbots) | Twitter killed free API in Feb 2023; app discontinued | Years of revenue from premium iOS client | Proprietary API = platform can kill you overnight |
| Twitterrific (Iconfactory) | Same API changes; app discontinued after 16 years | One of the original Twitter clients (est. 2007) | Decades of work destroyed by one API policy change |
| Apollo (Reddit) | Reddit introduced $20M/year API fees in Apr 2023; shut down Jun 30, 2023 | 1.5M monthly users; premium iOS client | Even beloved apps with loyal users can’t survive platform hostility |
| RIF / Sync / BaconReader | Same Reddit API pricing; all shut down Jun 30, 2023 | Multiple popular Reddit clients killed simultaneously | Platform risk is not hypothetical; it is inevitable |
| Hundreds of Twitter bots/tools | 2012 token limits, 2023 API paywall | Entire ecosystem of research tools, bots, analytics destroyed | Twitter’s API history is a decade-long series of broken promises |
Every casualty above was killed by the same mechanism: a single company changed its proprietary API terms. The AT Protocol is fundamentally different:
However: In practice, Bluesky still operates the dominant Relay and App View. True federation is still evolving. The “open protocol” protection is real in theory but untested at scale.
The global social media management market is valued at $30–36B (2025–2026), growing at 14–17% CAGR. The social media analytics market is $10–16B, growing at 24–27% CAGR. Even a tiny slice of these markets, focused on Bluesky’s 40M users, could sustain a bootstrapped business.
Growth has decelerated from 5M users/month (peak Nov 2024) to 1.6M/month (mid-2025). Post volume has declined ~66% from peak. Daily likes down ~64%. The platform adds users but many don’t stick. With 30-day retention at 18%, Bluesky has a “revolving door” problem.
Counter-argument: 40M registered users with ~4M DAU is still meaningful. The platform continues to grow at ~0.5 users/second. Event-driven growth spikes will continue as long as X remains controversial.
Meta’s Threads has 450M MAU and 137–150M DAU—more than 10x Bluesky on every metric. Threads doubled its user base in under a year (200M → 400M). It has Instagram’s built-in distribution and Meta’s ad infrastructure. For brands and mainstream users, Threads may be the “good enough” X alternative.
Counter-argument: Threads is a Meta product (centralized, algorithmic, data-harvesting). Bluesky appeals to a different audience: those who value openness, decentralization, and algorithmic choice. The two may coexist rather than compete directly.
Mastodon peaked at 2.6M MAU in November 2022, now below 690K—a 73% decline. The “decentralized social network” pitch wasn’t enough to retain mainstream users. Difficulty recreating social graphs, confusing server selection, and arbitrary moderation drove people away.
Counter-argument: AT Protocol is architecturally superior to ActivityPub for user experience. Global namespace (no server confusion), account portability, algorithmic choice, and a single-app UX that hides protocol complexity. Bluesky has already surpassed Mastodon’s peak by an order of magnitude.
Bluesky’s 2026 roadmap includes features that overlap with third-party tools: better discovery, post analytics (in Bluesky+ mockup), drafts (shipped), and real-time event features. As Bluesky grows its team (from 25 to 30+ employees), the feature surface will expand.
Counter-argument: With only ~6 engineers building app features, Bluesky can’t build everything. The AT Protocol is explicitly designed to delegate functionality to third parties. Bluesky has partnered with Buffer rather than building their own scheduling. The cultural commitment to ecosystem support is genuine (grants, conferences, Skyseed fund).
X still has 540–570M MAU. If Musk’s leadership stabilizes or improves, the exodus could slow. Most “news influencers” who joined Bluesky are still more active on X. X’s network effects are powerful.
AT Protocol is still evolving. OAuth support is just arriving. Custom schemas are still being defined. True federation (multiple independent Relays and App Views) is not fully realized. The IETF standardization process just started in Jan 2026.
Can a 40M-user platform with 4M DAU support a SaaS ecosystem? For context: social media management tools that survive tend to be multi-platform. Building on Bluesky alone limits your TAM. The “build for Bluesky + others” strategy is more defensible.
This is the most concerning signal. Despite adding users, Bluesky’s daily post volume fell from 1.48M (Nov 2024 peak) to ~500K (Jun 2025). A Pew Research Center analysis showed engagement fell ~50% from its peak. If the active core shrinks, the tooling market shrinks with it.
The Bluesky ecosystem values free, open-source tools. The most successful tools (ClearSky, SkyFeed, Blueview) are free. This means:
This is the most defensible approach. It solves the “Bluesky is too small” problem while capturing the “Bluesky is the most developer-friendly” advantage.
Bluesky’s culture is uniquely receptive to “building in public.” SaaS founders who build in public grow audiences 3x faster than those who don’t (per Hypefury). The platform values:
Content mix: 60% value-driven (tips, insights, UGC), 30% community engagement (replies, conversations, reposts), 10% direct promotion.
The Bluesky ecosystem is heavily open-source. The protocol itself is open. Developer grants require open-source. Consider open-core: free open-source tool + paid cloud hosting/premium features.
Bluesky thrives on community. Build your tool’s community ON Bluesky itself:
| Model | Example | Notes |
|---|---|---|
| Freemium | Buffer (free tier + $6–$120/mo) | Free tier with limited accounts/posts; paid for more capacity |
| Usage-based | Graze (30% of ad revenue) | Works for tools that generate revenue for users |
| Flat subscription | BlueSkyHunter ($15–$29/mo) | Simple and predictable for users |
| Per-seat | Sprout Social ($199–$399/seat/mo) | For team/enterprise tools; premature for Bluesky |
| Open-core | Plausible, PostHog model | Free self-hosted; paid cloud with extras |
| Metric | Estimate |
|---|---|
| Bluesky total users | ~40–43M (Feb 2026) |
| Daily active users | ~3.5–4.1M |
| Daily unique posters | ~1.3M |
| Active brands (est.) | <50K (very early) |
| Power users (est. posting daily, 5K+ followers) | ~50K–100K |
| Custom feeds | 50,000+ |
| Third-party dev tools/apps | Dozens (most free/hobby projects) |
| Paid Bluesky tools (est.) | <10 (BlueSkyHunter, BskyGrowth, Graze ads) |
| Breakeven target (solo founder) | ~300–500 paying users at $20/mo = $6K–$10K MRR |
| API/data cost | $0 (free firehose) + hosting for your service |
Building a SaaS on Bluesky in February 2026 is a calculated bet on growth. The platform has real users (40M+), a genuinely open API (free), a developer-friendly culture, and structural protection against platform risk (open protocol). The downside is clear: engagement is declining from peak, retention is weak (18% at 30 days), and Threads is growing 10x faster.
The winning formula: Build a multi-platform tool that is Bluesky-first but supports X, Threads, and Mastodon. Use the free Bluesky API to build your product with near-zero data costs. Market by building in public on Bluesky itself. Start free, convert to paid. If Bluesky grows, you’re perfectly positioned. If it doesn’t, your multi-platform support means you can pivot emphasis.
The worst approach: Build something that works ONLY on Bluesky with no path to other platforms. History teaches us that single-platform dependency is the #1 killer of social tools businesses—even when the platform is open.