AILIMIT.WATCH

Replit

Last verified: Jun 6, 2026

1 unannounced

12 sourced changes·58% official·verified against official docs, changelogs & Wayback·last checked Jun 6, 2026

01 / VERDICT

Stable

Low recent constraint volatility — few meaningful restrictions in the last 12 months.

WHY STABLE — trend: stable · no recent tightening

Constraint Stability

88/100

higher = fewer recent restrictions

stable0% / 90d
1 unannounced change

Replit has stayed stable, with minimal changes to its limits over time.

Most noticeable: Developers depending on the code-execution API lost it entirely

Recency
248
days since it last tightened
Velocity
1.5
tightenings per year
Silent
17%
of tightenings were unannounced
Reversals
50%
of tightenings later reversed (same dimension)

Replit constraint metrics — computed from 6 tracked tightenings · median 188 days between tightenings. Original to this site; not published by the vendor.

IMPACT BY PERSONA
Solo / occasional
Brief, occasional use is rarely touched by limit changes.
low
Power user / daily heavy
Daily-limit, session and rate constraints have been steady.
low
Team / shared workflow
Exposed to usage-based cost changes across the team.
medium
Startup / cost-scaling
Exposed to usage-based cost as usage scales.
medium
Enterprise / long-term dependency
Continuity exposure: ownership / policy shifts.
medium

02 / ANALYSIS

Replit carries one of the longer tightening records here, reaching back to 2022. The early moves stripped infrastructure guarantees: the code-execution API was retired in October 2022, leaving dependent developers with nothing, and outbound data-transfer limits began being enforced in April 2023. Pricing then consolidated — December 2023 folded the monthly Hacker ($7) and Pro ($20) tiers into an annual Core membership — before billing turned decisively usage-based. July 2025 swapped a flat $0.25/checkpoint for 'effort-based' variable pricing, and once Agent 3 launched in September 2025 its autonomy drove runaway costs: users reported $1,000 in a single week against typical $180–250 bills, with Replit itself conceding it 'can be more expensive over a project lifetime.' February 2026's overhaul made Core credits expire after six months and retired the Teams tier; a 24 February follow-up cut Core from $25 to $20 and added a $100 Pro plan, migrating Teams users across. Mid-2026 additions — a default-on Package Firewall, persistent Agent customization, a Databricks preview, Claude integration — lean neutral-to-loosening. The consistent motion is fixed, predictable cost giving way to variable, consumption-driven cost, with expiry quietly attached to credits already paid for. The real risk isn't losing access but losing cost predictability: effort-based pricing plus agentic autonomy can push a month's spend well past the plan price, which is why hard budget caps matter more here than the headline tier.

03 / CURRENT LIMITS

04 / WHERE IT MATTERS

Which usage patterns the recent changes affect — each line backed by the events driving it.

Agentic / high-volume costmedium exposure
Feb 2026↓Pricing overhaul; credits now expire; Teams retired·source
Sep 2025↓Agent 3 autonomy causes surprise cost overruns·source
Jul 2025↓Flat checkpoint price → variable effort-based cost·source
Dec 2023↓Monthly Hacker/Pro tiers folded into an annual Core membership·source
Oct 2022↓Developers depending on the code-execution API lost it entirely·source
High-frequency usagelow exposure
Apr 2023↓Hosted apps that exceed egress limits get throttled·source

05 / WHAT CHANGED

Technical 3 · Product 4 · Business 5 · Policy 0

  1. Dec 2023

    ↓ Monthly Hacker/Pro tiers folded into an annual Core membership

    businessImpact Medium · source · secondary source
  2. Apr 2023

    ↓ Hosted apps that exceed egress limits get throttled

    technicalImpact Medium · source · secondary source
  3. Oct 2022

    ↓ Developers depending on the code-execution API lost it entirely

    technicalImpact High · source · secondary source
▸ View full constraint history (9 events)
  1. Jun 2026

    ↑ Expands access to the Replit-Databricks integration from a limited release to a public preview.

    technicalImpact Low · source · secondary source
  2. Feb 2026

    ↑ Core price cut 20%; Teams migrated to Pro

    businessImpact Low · source
  3. Feb 2026

    ↓ Pricing overhaul; credits now expire; Teams retired

    businessImpact High · source · secondary source
  4. Sep 2025

    ↓ Agent 3 autonomy causes surprise cost overruns

    businessImpact High · unannounced · source

    ↔ shipped alongside Agent 3 released (Sep 2025) — more for more, not a pure cut.

  5. Sep 2025

    → Agent 3 released

    productImpact Low · source
  6. Jul 2025

    ↓ Flat checkpoint price → variable effort-based cost

    businessImpact Medium · source
  7. Dec 2023

    ↓ Monthly Hacker/Pro tiers folded into an annual Core membership

    businessImpact Medium · source · secondary source
  8. Apr 2023

    ↓ Hosted apps that exceed egress limits get throttled

    technicalImpact Medium · source · secondary source
  9. Oct 2022

    ↓ Developers depending on the code-execution API lost it entirely

    technicalImpact High · source · secondary source
▸ Other product updates (3) — feature launches, not constraint changes
  1. Jun 2026

    → Reduces friction by enabling direct workflow integration between Claude and Replit.

    productImpact Low · source · secondary source
  2. Jun 2026

    → Adds a security layer that blocks malicious packages during installation, potentially preventing some package installs.

    productImpact Low · source · secondary source
  3. Jun 2026

    ↑ Reduces repetitive prompting and context-setting for AI agents by allowing persistent custom instructions and skills.

    productImpact Low · source · secondary source

06 / OUTLOOK

Volatile

Volatile. Last 12 months: 2 pricing, 1 limit/quota, 3 model-availability changes (1 tightening).

Observed past cadence — not a prediction.

07 / USER SIGNALS

USER SIGNALS — HACKER NEWS · GITHUB · APP STORE

14 complaint mentions · last 180d

14 HN

  • Pricing·1071%
  • Model removal·214%
  • Limits / quota·17%
  • Performance·17%

Hacker News, GitHub issues & Apple App Store reviews matching complaint keywords — not a scientific sample. See methodology.

08 / ALTERNATIVES

ALTERNATIVES

By constraint history — not capability.

More stable:Gemini, Grok, v0
Similar risk:Tabnine, Perplexity, Claude, Devin
→ Full comparison matrix

09 / RELATED