← demos

Kalmari · Frontrun Bot · Market Research

Competitor & Market Research

Pure competitive landscape — who else builds trading bots/tools for Kalshi & Polymarket, and how each compares to Kalmari. Compiled 2026-06-17.

Why this document exists. The research/ folder had no standalone market-research artifact — only GTM strategy (GTM_PLAN.html), paid-ads strategy (KALMARI_GOOGLE_ADS_RESEARCH.md, whose §6 was an explicit summary pointing to a "full deep-dive" that was never written), a legal audit, and a code audit. This file is that missing deep-dive: a multi-source, adversarially fact-checked competitor sweep. Findings below survived 3-vote verification; unverified names are flagged, not asserted.

1. Executive summary & wedge verdict

Kalmari's defining wedge is the four-way intersection of (1) Telegram-native, (2) US-regulated Kalshi surface, (3) autonomous AI-strategy execution, and (4) BYOK / non-custodial. The verified field splits into two camps that each cover three of these four axes but miss the fourth — leaving the full intersection effectively unoccupied as a managed commercial product.

Verdict: No competitor occupies all four axes simultaneously as a product. The closest managed threats are Camp A (everything but US-regulation) and Camp B CLIs (everything but Telegram-native). Kalmari's specific slot — a hosted, no-self-deploy Telegram bot on US-regulated Kalshi with AI-strategy (not copy-trade) execution and free unlimited paper trading — is genuinely empty in the verified data.

Single largest risk to this verdict. Polymarket is standing up a CFTC-regulated US arm (QCEX / QCX DCM; waitlist reportedly removed May 2026). If Camp A bots (KreoPoly / PolyGun / PolyBot) integrate that regulated arm with fiat rails, they would close Kalmari's US-regulated axis and invade the wedge directly. This is fast-moving and must be re-checked.

2. Master comparison matrix

Kalmari row highlighted. "Autonomous" = places live trades by itself (vs signals-only). "BYOK" = non-custodial, user supplies own keys. Pricing left blank where it could not be verified from primary sources.

Product Surface Markets Autonomous? Non-custodial / BYOK US-regulated Pricing / free tier Traction
Kalmari Telegram bot (managed) Kalshi only Yes (AI strategy) Yes Yes Free unlimited paper · paid live (TBD) Pre-launch
KreoPoly Telegram bot Polymarket (Kalshi secondary) Yes (copy-trade) Yes (Privy/Safe) No (crypto) Not verified (likely per-trade) Heavy press Feb–Mar 2026
PolyGun Telegram bot Polymarket Yes (copy/sniper) Yes No (crypto) Not verified Active
PolyBot Telegram bot Polymarket only Yes (rules engine) Yes (Gnosis Safe) No (crypto) Not verified · $1 min deposit Active
ryanfrigo/kalshi-ai-trading-bot CLI + Streamlit (OSS) Kalshi only Yes (--live) Yes Yes Free (open-source) GitHub repo
OctagonAI/kalshi-trading-bot-cli CLI (OSS, Bun) Kalshi only Partial* Yes Yes Free (open-source) GitHub repo
yllvar/Kalshi-Quant-TeleBot Telegram bot (OSS, self-host) Kalshi only Claimed Yes Yes Free (MIT) Unknown / unmaintained
ImMike/polymarket-arbitrage Web dashboard (OSS, FastAPI) Polymarket + Kalshi Yes (arb)* Yes No (Polymarket-first) Free (open-source) Skeleton (~4 commits)
Alphascope Web app Polymarket, Kalshi, Manifold, Opinion No (signals only) n/a n/a (info tool) Free tier Claims 10,000+ traders
OddsJam Web/mobile Sportsbooks (not pred-mkts) No (you place bets) n/a n/a Subscription Established
Composer Web/mobile Stocks/ETF/crypto/options Yes (out of category) Custodial (Alpaca) n/a Subscription Established

* OctagonAI: its claim to generate its own probability estimates and place edge-vs-orderbook orders was refuted (1-2) in verification — autonomy depth unconfirmed. ImMike: autonomous-execution detail passed only 2-1; repo is simulation-focused.

3. Competitor profiles by tier

Tier 1 — Direct threats (Telegram-native execution bots)

KreoPoly crypto / Polymarket-first

Telegram bot · Polymarket primary, Kalshi secondary · non-custodial (Privy enclaves + Gnosis Safe) · copy-trading

The single closest direct threat. A non-custodial Telegram bot that autonomously copy-trades prediction markets — mirroring whale wallet trades onto the user's own wallet in real time. Matches Kalmari on 3 axes (Telegram-native, autonomous, non-custodial). Mechanism differs: copy-trading whales vs Kalmari's AI-analyzed strategy signals.

Strengths vs Kalmari
  • Live product, heavy press coverage (Feb–Mar 2026)
  • Multi-platform (Polymarket + Kalshi integration)
  • Real-time copy-trading of proven wallets
Weaknesses vs Kalmari
  • Polymarket-primary = crypto/offshore, legally shaky for US retail
  • Copy-trade only — no owned, named AI strategy
  • "BYOK" is loose (Privy generates keys = managed self-custody)

PolyGun crypto / Polymarket-first

Telegram bot · Polymarket · non-custodial · copy + sniper orders

A true platform-surface analog to Kalmari: browse, market/limit/sniper orders, portfolio tracking, and autonomous copy-trading entirely inside Telegram. But crypto-first — deposits via autobridging from Polygon, Solana, Ethereum, BNB on the offshore Polymarket surface (CFTC-restricted for US retail). Matches Telegram-native + autonomous + non-custodial; misses US-regulated.

PolyBot crypto / Polymarket-first

Telegram bot · Polymarket only · non-custodial (per-user Gnosis Safe, $1 min) · rules engine

The most feature-rich Polymarket bot: time-windowed strategies (5m/15m/1h/4h/24h), an Auto Trader rules engine (entry/exit, take-profit, stop-loss, trailing stops), and leaderboard copy-trading. Self-custodial via per-user Gnosis Safe on Polygon, exportable keys. Matches 3 axes; Polymarket/Polygon only, no Kalshi — misses US-regulated.

Tier 2 — Open-source bots (lead-gen, not managed competitors)

ryanfrigo/kalshi-ai-trading-bot US-regulated (Kalshi)

CLI + optional Streamlit · Kalshi only · BYOK (Kalshi API key + RSA key) · paper mode · --live flag

The most architecturally relevant open-source tool. AI-automated trading on Kalshi via BYOK, with paper-trading default and live orders gated behind --live. Critically, it ships a "Safe Compounder" strategy (NO-side resting maker orders, edge/EV math, no LLM) that closely parallels Kalmari's "Safe NO Bets." Matches US-regulated Kalshi + autonomous + BYOK + strategy design; misses Telegram-native, and autonomy is operator-run CLI (cron/manual), not hosted.

OctagonAI/kalshi-trading-bot-cli US-regulated (Kalshi)

CLI (Bun runtime) · Kalshi only · BYOK (keys local in ~/.kalshi-bot/.env) · Telegram = outbound alerts only

An AI-native command-line tool (the seed-list name "kalshi-deep-trading-bot" is a rename of this repo). Polymarket appears only as a GitHub topic tag — no actual support. Has outbound Telegram trade alerts, but that's a notification sink, not a chat interface. Matches US-regulated Kalshi + BYOK; misses Telegram-native. Its claim to compute its own probabilities and place edge-vs-orderbook orders was refuted in verification — treat autonomy as unconfirmed.

yllvar/Kalshi-Quant-TeleBot US-regulated (Kalshi)  ← closest full-wedge analog

Telegram bot (JS, self-host, MIT) · Kalshi only · BYOK (Kalshi API key + Telegram bot token via env)

The only verified competitor that hits all four axes: Telegram-native + US-regulated Kalshi + autonomous + BYOK. The catch: it is unmaintained self-host code the user must deploy themselves — not a managed product — and its "autonomous/professional-grade" framing rests on README self-description (headline claim passed only 2-1). Maturity, maintenance, and real usage are unknown. This is the name to keep watching: if someone wraps it into a hosted service, the wedge narrows.

ImMike/polymarket-arbitrage Polymarket-first

Web dashboard (Python, FastAPI) · Polymarket + Kalshi · BYOK · simulation-focused

An open-source arbitrage bot monitoring both Polymarket and Kalshi (self-claims 10,000+ markets) for cross-platform and intra-platform (YES+NO bundle) inefficiencies. Auto-submits detected signals to an execution engine with dry-run default + live mode. No Telegram — monitoring via local web dashboard. Low-maturity skeleton (~4 commits). Different surface (web) and different job (arbitrage, not strategy) from Kalmari.

Tier 3 — Analytics & tangential EV tools (out of category)

Alphascope multi-market analytics

Web app · Polymarket, Kalshi, Manifold, Opinion · signals/research only

A multi-market signals/research web dashboard (Forecast/Predictions/Odds/News) — not a Telegram bot and not an execution engine. Explicitly states the trader makes the decision: "Alphascope cannot make those decisions for the user." Claims 10,000+ traders; has a Telegram community channel but no trading bot. The category SEO leader. Kalmari's wedge against it: "Alphascope tells you; Kalmari does it."

OddsJam & Composer out of category

OddsJam: positive-EV sports-betting tool (you place every bet yourself) · Composer: stocks/ETF/crypto/options algo-trading via Alpaca (custodial)

OddsJam is a signals-only positive-EV/arbitrage sportsbook tool — it tells you the book, game, market and stake, but you place the bet; only automation is tracking/grading. Composer trades stocks/ETFs/crypto/options, not prediction markets, with no evidence of a Kalshi pivot as of 2026. Both are reference points, not competitors.

4. Regulatory split

The clearest dividing line in the market is regulatory posture — and it maps almost exactly onto which platform a product targets.

US-regulated (Kalshi / CFTC DCM)Crypto / Polymarket-first (legally shaky for US retail)
Kalmari, ryanfrigo/kalshi-ai-trading-bot, OctagonAI/kalshi-trading-bot-cli, yllvar/Kalshi-Quant-TeleBot KreoPoly, PolyGun, PolyBot, ImMike/polymarket-arbitrage

Every managed Telegram product in the field (KreoPoly/PolyGun/PolyBot) sits on the crypto side. Every US-regulated tool is open-source/self-deploy. Kalmari is the only entrant proposing a managed product on the regulated side — that is the wedge in one sentence.

5. Pricing-band analysis

Implication for Kalmari: there is no incumbent subscription price to anchor against in the regulated-Kalshi-bot category — it's empty. Kalmari can set the reference price. A free unlimited paper tier → flat monthly live model is differentiated against both the free-but-self-deploy OSS tools and the fee-on-trade crypto bots.

6. Positioning gaps nobody is claiming

  1. Managed (hosted, no self-deploy) Telegram bot on US-regulated Kalshi — the exact empty slot. The only Telegram+Kalshi+autonomous code (yllvar) is self-host OSS; everything hosted is crypto.
  2. AI-strategy execution, not copy-trading. Every managed Telegram bot in the field is a copy-trade/mirror bot. Kalmari's owned, named strategies ("Safe NO Bets," "Rapid Resolution") are a distinct, clearer pitch than "mirror this whale."
  3. Free unlimited paper trading. OSS tools require deployment to paper-trade; copy bots are live-money-first. Unlimited managed sim ("simulate forever, go live when ready") is unclaimed.
  4. Public, verifiable live track record. No competitor — open-source or commercial — publishes an auditable live P&L. This alone is a credibility multiplier.
  5. BYOK non-custodial safety story on the regulated surface. Camp A's "non-custodial" is crypto-wallet non-custodial; Kalmari's is regulated-brokerage BYOK — a cleaner safety narrative for US retail, especially after any TG-bot rug event.

7. Caveats & unverified names

Did NOT survive verification (existence / positioning unconfirmed — not asserted above): okbet, KalshiAI, Rainmaker, Polytrader, Polybro, Astron, Billy Bets, Pigeon, Polysights, Verso, TradeFox, Betmoar, Polycule, Polymtrade, Stand, Dome, Adjacent News, RebelBetting, 3Commas, TrendSpider.
These appeared in the seed list (largely from the prior ads-doc §6 summary) but could not be confirmed as Telegram+Kalshi execution bots from primary sources in this pass. Several may exist in adjacent forms (e.g. RebelBetting/3Commas are real products but out of category); others may be defunct, renamed, or mis-attributed. Worth a targeted re-check before citing any as a live competitor.

8. Open questions to close next

9. Sources

Primary sources prioritized (product sites, pricing pages, GitHub repos). 23 sources fetched · 102 claims extracted · 25 verified · 24 confirmed · 1 killed · 11 findings after synthesis.