Show DEV: Building an Anonymous Instagram Profile & Story Viewer Without Logins

x z — Fri, 25 Sep 2026 08:29:23 +0000

Hey DEV Community! 👋

I recently launched a lightweight side project called IGViewer and wanted to share the motivation, technical hurdles, and architecture behind it.

💡 Why Build This?

Instagram increasingly walls off public content behind forced login prompts. If you just want to:

Quickly inspect a public competitor's campaign without logging into your brand/work account
Check a public story without leaving your personal username on the official 24-hour viewer list
Collect reference materials or public post grids for research

...you are usually forced into either logging in or using clunky, ad-infested third-party tools that break after two clicks.

I wanted to build a clean, fast, and transparent browser-based tool: IGViewer.

🚀 What It Does

Anonymous Active Stories: View 24-hour public stories without appearing in the account's viewer list.
Profile & Grid Previews: Check bios, profile pictures, post feeds, and reels previews without logging in.
Highlights & Downloader: Browse public highlights and save original-resolution media.
Strict Privacy Boundaries: Public content only. No private profile scraping, no DM tools, and never asking for user credentials.

🛠️ Technical Challenges & Lessons Learned

Building an interface on top of heavily rate-limited and ephemeral platforms brought a few interesting engineering challenges:

Handling Ephemeral Media (Stories):

Stories expire in 24 hours. Managing accurate cache expiration without serving stale media or hammering downstream endpoints required strict TTL caching at the edge.
Rate Limiting & Cost Management:

To prevent abuse and bot spam while keeping the service sustainable, I implemented a tiered quota system:
- Free daily public lookups without registration.
- Credit-based unlock path for deeper pagination, highlight playback, and follower lists.
Client-Side Performance:

Instagram profiles have heavy asset payloads (thumbnails, video previews). We lazy-load media grids and proxy/compress preview assets so the page loads in under a second even on mobile connections.

🔗 Try It Out & Feedback

You can test the tool live here: https://www.igrecent.com/

I’d love to hear feedback from fellow builders:

How would you optimize the caching layer for ephemeral social media content?
Any UI/UX suggestions for the profile navigation?

Thanks for reading!

Fast Decisions in Agent Workflows: Laya vs TypeSafe Jev

x z — Tue, 22 Sep 2026 04:27:41 +0000

Fast Decisions in Agent Workflows: Laya vs TypeSafe Jev

Originally published on JevLab. This is an independent, unofficial evaluation write-up — not affiliated with TypeSafe or Convai Innovations.

TL;DR: Using a 70B autoregressive LLM for simple agent routing is increasingly the wrong default. Convai Innovations' open-source Laya (421M ModernBERT-large) challenges TypeSafe's hosted Jev primitive. Published charts put Laya roughly 7.8× faster (32.8 ms vs 236–276 ms), with tighter calibration after temperature refit (ECE 0.081 vs 0.246), plus broad multilingual reach at self-hosted software cost of $0. But in production agent pipelines, raw accuracy is only half the story: confidence thresholds and fail-closed handoffs decide whether a wrong route ever executes.

1. The "breakthrough" debate and System 1 decision engines

In early September 2026, TypeSafe launched its proprietary Jev engine (e.g. jev-1.13.0) and framed non-autoregressive decision primitives as a breakthrough for AI workflows. Instead of generating variable-length text, Jev returns a deterministic choice plus a confidence score.

Shortly afterward, Nandha Kishor M (Convai Innovations) published a widely shared DEV post: I Built Non-Autoregressive Decision Models a Year Ago. Then a Frontier Lab Called It a "Breakthrough". Convai also released Laya — Apache-2.0, ModernBERT-large weights on Hugging Face (convaiinnovations/laya).

(If you have the exact DEV URL for Nandha's post, swap it into the link above before publishing.)

This is an architectural fork in agent engineering: System 1 (fast, non-generative classification) vs System 2 (slow, autoregressive reasoning).

Incoming request
        │
[System 1: fast decision engine]
(Laya ~33ms local / Jev hosted API)
        │
Confidence >= threshold?
   /              \
 YES               NO
  │                 │
Auto-adopt route   Fail-closed handoff
(specialist tool)  (human / fallback)

Using GPT-4 / Claude-class models just to pick among five tools often costs hundreds of milliseconds to multiple seconds, burns tokens, and risks schema/JSON failures. Decision engines skip text decoding: one forward pass → categorical log-probabilities.

2. Head-to-head numbers (with an important caveat)

Convai published a multi-axis comparison of Laya against third-party published TypeSafe Jev figures (e.g. work attributed to AbdelStark and nlbzard). That means useful directional signal — not a byte-identical side-by-side under identical network conditions.

Metric	TypeSafe Jev (1.13.0 published)	Laya (Convai)	Delta
P50 latency (1 q)	236–276 ms (hosted API)	32.8 ms (local T4)	~7.8× faster
Batched (50 q)	N/A (concurrency capped)	7.2 ms / q (337 ms total)	local scale
Calibration (mean ECE)	0.246	0.081 (temp refit)	~3× tighter
License / price	Closed API (~$0.042 / 1M tokens)	Apache 2.0 ($0 self-hosted software)	open
Multilingual	Unspecified / EN-focused	45 / 51 languages usable	broad
Privacy	Public endpoint egress	Air-gapped / local	zero egress

Accuracy on public sets (as reported)

typed-decisions (2,000 evals): Laya 0.766 vs Jev 0.727 (+3.9%); Laya above estimated teacher ceiling (~0.735)
AG News (4 labels): 0.950 vs 0.910 (+4.0%)
DAIR Emotion (6 labels): 0.595 vs 0.480 (+11.5%)

Held-out workflow-style tasks (Laya checkpoints)

Phishing: 0.940–0.993
Email spam: 0.958–0.993
Topic categorization: 0.930–0.953
Jailbreak / guardrails: 0.708–0.762
RAG passage relevance: 0.625–0.657
10-way support triage: 0.502–0.522

3. Why Laya can be ~33 ms: ModernBERT + calibration training

ModernBERT-large (421M)

Encoder-style stack (FlashAttention-2, RoPE, unpadding, larger context) evaluates tokens in parallel. That avoids the O(N) decode loop of autoregressive routers.

RLCD-style calibrated training

Classic classifiers overfit to top-1 argmax and become overconfident. Laya-style training pushes toward proper scoring (Brier / log-loss), so probabilities are more usable as policy inputs, not just leaderboard cosmetics.

Checkpoint specialization + preload

from laya import Router

router = Router(preload=True)

decision = router.route({
    "task": "Review pull request #104 for potential race conditions",
    "context": "diff --git a/worker.go b/worker.go..."
})

Mixed-language workloads that reload weights pay a large cold-load penalty (Convai cites ~7.4s). Preloading keeps per-call latency flatter across languages.

4. The part that actually ships: ECE + fail-closed routing

In agents, calibration beats raw accuracy.

Predict code_search at 0.98 → auto-adopt is reasonable.
Predict code_search at 0.52 → you are guessing; fail-closed handoff (human review or a clarifying question) is the safe policy.

ECE (lower is better)

Laya (shipped raw):     ~0.466
Jev (published):        ~0.246
Laya-Multi (refit):     ~0.106
Laya (temp refit):      ~0.081

Uncalibrated models can say "95% confident" on inputs they get wrong ~40% of the time. Temperature / domain refit is not optional trivia — it is how you make thresholds meaningful.

5. JevLab's angle: measure the handoff boundary

Our working thesis:

Accuracy without thresholding is an illusion. In production, your policy lives or dies by the handoff boundary.

Two caveats we keep repeating:

Third-party tables ≠ controlled A/B. Prefer identical prompts, identical network/runtime assumptions, and pinned model versions.
Ambiguity is the real production tax.
- "Find where the auth token is parsed and update the tests to mock it" → multi-intent (code_search + test_runner)
- "What does this function do?" with no code attached → information-deficient

An unconstrained router will still emit a label with inflated confidence and corrupt downstream state. Thresholding exists to turn "I'm guessing" into handoff, not into a wrong tool call.

How we structure evals conceptually:

Clear validation/test sets → auto-adopt rate and error among auto-adopted decisions
Adversarial / ambiguous challenge sets → does confidence collapse so a threshold like t = 0.80 can trigger handoff cleanly?

6. Production decision matrix

Prefer Laya if you need:

sub-~50 ms P50 locally
air-gap / no egress
existing GPU (T4/L4-class) and ops willingness
strong multilingual coverage

Prefer TypeSafe Jev if you need:

serverless / zero GPU ops
pure TypeScript / edge-friendly integration
low or bursty volume where API cost beats a always-on GPU

Need zero egress / HIPAA / air-gap?
        /                \
      YES                 NO
       │                  │
    Choose LAYA     Have GPU infra?
                     /          \
                   YES           NO
                    │             │
              Need <50ms?     Choose JEV
               /      \       (serverless)
             YES       NO
              │         │
          LAYA     Evaluate both

7. What we're building next

Open decision models like Laya reinforce the bet: System 1 routing is infrastructure, not a demo.

We're evaluating a Laya baseline inside JevLab so builders can:

Compare engines on the same agent-routing tasks
Sweep confidence thresholds and watch auto-adopt vs handoff trade-offs
Export TypeScript-oriented router policy shaped by those curves

Until a published measured report lands, treat the public site as an unofficial threshold lab / preview — useful for thinking in policy space, not as a claim that every chart is a finalized ship gate.

Read the full original + try the lab: https://jevlab.dev/blog/laya-vs-jev · https://jevlab.dev

Unofficial independent project. Not affiliated with, endorsed by, or maintained by TypeSafe AI or Convai Innovations. Benchmark figures cited above come from published third-party / vendor materials unless otherwise noted.

I was tired of sign-up walls for AI video tools, so I built my own browser app

x z — Tue, 28 Jul 2026 12:32:50 +0000

Every time I wanted to quickly test a prompt or generate a short video clip with models like Kling or VEO, I kept running into mandatory sign-ups, paywalls, or super complex node-based UIs just for a simple test.

I wanted something clean, fast, and friction-free. So I built RenderPop.

What is RenderPop?

It's a lightweight browser workspace for AI image and video generation.

Instead of jumping between 5 different tabs for different models, I aggregated top-tier image and video models under one roof.

What you can do with it:

Instant Fast Generation: Try out fast image generation right away without forcing an account registration.
Multi-Model Access: Access top image and video models in a single interface.
Motion Control & Image-to-Video: Animate still portraits or apply dance motion templates directly onto reference photos.

Feedback Welcome!

I'm constantly tweaking the interface and performance. If you give it a spin at RenderPop, I'd really appreciate your thoughts on the UI or any features you think I should add next!

DEV Community: x z

Show DEV: Building an Anonymous Instagram Profile & Story Viewer Without Logins

💡 Why Build This?

🚀 What It Does

🛠️ Technical Challenges & Lessons Learned

🔗 Try It Out & Feedback

Fast Decisions in Agent Workflows: Laya vs TypeSafe Jev

Fast Decisions in Agent Workflows: Laya vs TypeSafe Jev

1. The "breakthrough" debate and System 1 decision engines

2. Head-to-head numbers (with an important caveat)

Accuracy on public sets (as reported)

Held-out workflow-style tasks (Laya checkpoints)

3. Why Laya can be ~33 ms: ModernBERT + calibration training

ModernBERT-large (421M)

RLCD-style calibrated training

Checkpoint specialization + preload

4. The part that actually ships: ECE + fail-closed routing

5. JevLab's angle: measure the handoff boundary

6. Production decision matrix

7. What we're building next

I was tired of sign-up walls for AI video tools, so I built my own browser app

What is RenderPop?

What you can do with it:

Feedback Welcome!