adhd
Parallel divergent ideation method for LLM coding agents to escape premature convergence
ADHD delivers measurably more novel and broad ideas for open-ended coding problems where premature convergence is the enemy. If your agent needs to escape the obvious, this method works. Skip it for tasks with a single verifiable answer or when latency/cost are primary concerns.
Verified 1d ago · liveness 25/100 · cite: rightaichoice.com/tools/adhd
- Coding agents needing creative solution ideation
- Debugging fuzzy intermittent failures
- Interdisciplinary design tasks requiring non-obvious options
- Open-ended engineering problems with no single ground truth
- Tasks with a single verifiable answer
- High-latency or cost-sensitive applications
- Closed-form tasks where the textbook answer suffices
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip ADHD if you need a single correct answer, have strict latency or cost constraints, or if your environment lacks the Claude or Codex Agent SDK.
LLM call multiplication: N branches plus a critic pass per ideation, so API costs scale linearly with the number of branches you use.
ADHD is free and open-source, making it accessible to any developer with the required SDKs. Compared to commercial ideation tools, it has no subscription fees, but you incur your own LLM API costs, which can be higher than single-shot methods due to the branching and critic passes.
In short
adhd — Parallel divergent ideation method for LLM coding agents to escape premature convergence. Best for Coding agents needing creative solution ideation, Debugging fuzzy intermittent failures, Interdisciplinary design tasks requiring non-obvious options. Free to use.
What people actually say about adhd — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
91 mentions across 5 sources (Hacker News, YouTube, Bluesky, GitHub, Lemmy) · researched Jul 26, 2026.
- +Novel generator-critic separation prevents premature convergence.
- +Cognitive frames produce genuinely diverse solution approaches.
- +Open-source and free with no licensing restrictions.
- +Evaluated against single-shot baselines with measurable improvements.
- +15 built-in cognitive frames cover many design perspectives.
- −No partial-failure resilience — entire run aborts on single LLM error.
- −Wildcard frame guarantee breaks with '--frames 1' flag.
- −Works only with Claude & Codex Agent SDK, no Copilot support.
- −Limited real-world testing beyond six benchmark problems.
- −Scant user feedback — most data is off-topic about clinical ADHD.
- • Requires Claude & Codex Agent SDK usage, which may have its own API costs
- • Cloud compute costs if used at scale
Viability Score
How well maintained and how widely used is adhd? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- Parallel divergent ideation under cognitive frames
- Mechanical generator-critic separation via separate LLM calls
- No cross-branch context sharing during divergence
- Scoring, clustering, and deepening of top-K survivors
- Library of 15 cognitive frames (regulator, speedrunner, biology, etc.)
- Vantage-point reframing for branch diversity
- Generator system prompt forbids evaluation
- Single critic pass for convergence
- Prevents premature convergence
- Supports Claude and Codex Agent SDK
- Open-source method (preprint v0.1)
- Evaluated on six open-ended engineering problems
- Independent LLM-as-judge evaluation
About adhd
ADHD (Parallel Divergent Ideation) is an experimental open-source method for LLM-based coding agents that targets premature convergence—when a model latches onto the first plausible answer and polishes it, producing competent but forgettable output. Instead of a single linear trace, ADHD fans out N parallel branches, each under a structurally different cognitive frame from a library of 15 (e.g., regulator, speedrunner, biology, $0 budget). These branches generate independently with no cross-branch context, forcing true divergence rather than shared anchoring. A separate critic pass then scores, clusters, and deepens only the top-K survivors, enforcing a mechanical generator-critic separation via distinct LLM calls with opposing system prompts. The method is designed for open-ended engineering problems where the deliverable is a range of viable, non-obvious options—architecture decisions, API/SDK design, debugging fuzzy intermittent failures, refactor planning, and interdisciplinary tasks. In evaluations across six open-ended problems judged by an independent LLM-as-judge, ADHD outperformed a single-shot baseline on 5/6 tasks, with mean improvements of +5.17 in novelty, +4.17 in breadth, and +7.67 in trap detection (0–10 rubric). It avoids the convergence traps of Chain-of-Thought, Tree-of-Thought, and multi-agent debate, which all optimize for correctness on closed answer spaces. ADHD is available as a preprint (v0.1, 2026-05-25) and open-source implementation on GitHub. It requires the Claude and Codex Agent SDK to run. The method is ideal for coding agents that need to escape the obvious, but it is not suited for tasks with a single verifiable answer or for latency/cost-sensitive environments, as the parallel branching and critic passes add overhead. Compared to alternatives like ToT, ADHD's branches are isolated rather than shared, branching is driven by vantage-point reframing rather than next-step variation, and the generator/critic split is mechanically
Behind the Verdict
ADHD is a research method, not a polished product. It's a preprint and open-source implementation that targets a genuine problem: LLM agents converging on the obvious answer. The method's strength is its mechanical separation of generation and criticism, using separate LLM calls with opposing system prompts. This forces true divergence, as branches don't share context. The 15 cognitive frames provide diverse vantage points. In evaluations, it beat single-shot baselines on novelty, breadth, and trap detection. However, it's not for everyone. It multiplies LLM calls, increasing latency and cost. It's overkill for tasks with a single correct answer. It requires Claude or Codex Agent SDK. For teams building coding agents that need creative ideation, it's a valuable technique. For production environments where speed and cost are critical, it may be impractical. Compared to alternatives like Tree-of-Thought, ADHD's branches are isolated and driven by reframing, which is a key differentiator.
Researching adhd? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas adhd actually fits — and what changes day-one when you adopt it.
You need a range of architecture options beyond the typical layered pattern. You feed your requirements to an agent using ADHD, which generates multiple radically different options under frames like 'regulator' and '$0 budget'.
Outcome: You get several non-obvious architecture designs, each with distinct trade-offs, helping you choose a solution that fits your constraints and avoids the obvious trap.
The bug is fuzzy and hard to reproduce. You use ADHD to explore multiple causal frames, such as 'biology' and 'speedrunner', generating hypotheses that a single-shot approach would miss.
Outcome: You uncover a novel root cause hypothesis that leads to a fix, saving hours of trial and error.
You want to design an API surface that stands out. ADHD contrasts design philosophies from different frames, producing options that range from minimal to maximalist.
Outcome: You get a set of API design options with clear trade-offs, helping you pick one that aligns with your users' needs and your product vision.
Use Cases
- Generate diverse architectural options for a new microservice
- Debug an intermittent race condition by exploring causal frames
- Plan a large refactor with multiple non-obvious migration paths
- Brainstorm naming conventions that avoid corporate clichés
- Design an API surface by contrasting design philosophies
- Generate edge-case test scenarios beyond typical coverage
Models Under the Hood
as of 2026-08-10
Limitations
- ADHD is a preprint (v0.1) and open-source method, not a commercial product.
- It multiplies LLM calls per ideation (N branches plus a critic pass), increasing latency and cost.
- Evaluation uses an independent LLM-as-judge, which may introduce its own biases.
as of 2026-08-14
Verification history
We have re-verified adhd 9 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 9 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where adhd's pricing actually pencils out — and where peers do it cheaper.
ADHD is free and open-source, making it accessible to any developer with the required SDKs. Compared to commercial ideation tools, it has no subscription fees, but you incur your own LLM API costs, which can be higher than single-shot methods due to the branching and critic passes.
Setup time & first value
How long it actually takes to get something useful out of adhd — broken out by persona, not the marketing-page minute.
ADHD setup takes about an hour if you're familiar with the Claude and Codex Agent SDK. You'll need to clone the GitHub repo, install dependencies, and configure your API keys. The preprint provides guidance on implementation.
Integrations
Resources & Guides
Tutorials & Learning
Official links
Tools that pair well with adhd
Common stack mates teams adopt alongside adhd, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Adhd vs Praktika
These tools serve entirely different purposes: ADHD is a niche technique for coding agents to generate creative solutions, while Praktika is a mobile app for language learning. Choose ADHD if you're a developer building LLM agents that need to avoid premature convergence on open-ended problems. Choose Praktika if you're a language learner wanting to practice speaking with AI tutors. They are not competitors.
Adhd vs Temporal Ai
ADHD is a zero-cost, research-backed method for coding agents that need creative divergence and novelty, but it requires the Claude/Codex stack and isn't a product you can deploy. Temporal AI is a full-featured durable execution platform for building resilient, long-running AI workflows — if your need is reliability at scale, choose Temporal; if you want to boost agent creativity in open-ended coding tasks, try ADHD.
Adhd vs Audioeye
These tools are entirely incomparable in purpose: ADHD is a free ideation method for coding agents to generate creative solutions, while AudioEye is a paid enterprise accessibility compliance platform. Choose ADHD if you need novel, diverse ideas in engineering tasks; choose AudioEye if you need ADA/WCAG compliance with legal support.
Adhd vs Appgyver
Choose ADHD if you need a free, open-source method to boost creative coding agent ideation and avoid premature convergence on a single solution. Choose AppGyver if you are already an SAP customer and need a unified low-code/pro-code platform for building extensions, automating workflows, and integrating AI agents into your SAP landscape. These tools serve completely different use cases and hardly compete.
Adhd vs Cognition Ai
If you're a solo developer or researcher tackling open-ended design problems where you need creative, non-obvious solutions, ADHD's free, open-source method is a great fit. But if you're on an enterprise team shipping production code and need reliable, auditable automation with vendor support, Cognition AI's Devin platform—with FedRAMP compliance and a $10M productivity guarantee—is the clear choice despite likely higher cost.
Alternatives to adhd
View allFrequently Asked Questions
Best-of guides
Used adhd? Help shape our editorial sentiment research.