RightNow AI
RightNow Agent is a Rust terminal coding agent with signed native installers and a RunInfra coding plan from $10/mo.
The thing that kept us from recommending RightNow Agent — no signed installer for Windows and Mac — died on 2026-09-09 with release 1.0.4. What's left is a Rust harness that opens sessions in about 105 ms and a GPU kernel workflow (CUDA, Triton, PyTorch, Numba across 86+ architectures) that general CLI agents don't touch. Just be clear-eyed: the 5x-over-torch.compile numbers are vendor measurements, and you're paying separately for model tokens.
Verified 4d ago · liveness 82/100 · cite: rightaichoice.com/tools/rightnow-ai
- Developers who work primarily in the terminal and want a Rust CLI harness
- GPU kernel engineers doing CUDA, Triton, PyTorch or Numba optimization work
- Teams that want to bring their own model provider key instead of a bundled subscription
- Scripted and CI workflows that need headless one-shot runs with restricted tool sets
- Anyone who wants a GUI or IDE-style coding assistant instead of a CLI agent
- Air-gapped environments needing a local model — the agent doesn't install or run one
- Workflows that depend on image or video generation inside the agent
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 RightNow Agent if you need a GUI or IDE-style assistant, want image or video generation inside the agent profile, or need more than the fixed 1M token context any plan provides.
Your coding plan is shared by the whole workspace rather than per member, so every teammate draws from the same 5-hour and weekly limit pool.
The coding plans run $10/mo Starter (about $12 of usage), $29/mo Pro (about $35) and $99/mo Team (about $119), with self-serve credit top-ups from $5 for Model APIs. Team is shared per workspace, not per seat, so a small group is cheap relative to per-seat terminal agents. Below that, RunInfra Model APIs price per token — Nemotron 3.5 Lightning 30B at $0.05 input/$0.15 output per 1M tokens is the cheapest listed model.
In short
RightNow AI — RightNow Agent is a Rust terminal coding agent with signed native installers and a RunInfra coding plan from $10/mo. Best for Developers who work primarily in the terminal and want a Rust CLI harness, GPU kernel engineers doing CUDA, Triton, PyTorch or Numba optimization work, Teams that want to bring their own model provider key instead of a bundled subscription. Free to start; paid plans from $5.
What's new in RightNow AI
Checked 3 days agoAcross the latest 3 updates: 1 launch and 2 changelog entries.
RightNow Agent 1.0.4 ships native installers and in-app updates for six platforms
First hosted release adds native binaries for Windows, macOS and Linux plus install.ps1/install.sh, SHA-256 sidecars, GPG-signed Linux builds, and a verified in-app updater.
RightNow Agent 1.0.5 fixes Windows console pastes and model list refresh
RightNow Agent 1.0.5 folds multi-line Windows console pastes into one chip, refreshes /model after connect, and aligns Windows ARM64 artifacts with the signed binary.
RightNow Agent 1.0.3 legacy Windows x64 zip release
Legacy Windows x64 zip installs under %LOCALAPPDATA%\Programs\RightNow; superseded by the native installer in 1.0.4 with legacy files left intact.
What people actually say about RightNow AI — is it worth it?
We scanned public community sources for RightNow AI on Jul 3, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Our own analysis of that scan says the posts were off-subject. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
Viability Score
How well maintained and how widely used is RightNow AI? 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: October 2026
How we score →Key Features
- Terminal coding agent written in Rust, forked from xAI's grok-build
- Warm agent service opens sessions in a measured 105-115 ms
- Dynamic tool sending cut prompt tokens from 9,676 to 1,976 in recorded runs
- One-call transactions complete bounded create, edit or named-test from the tool result
- Native binaries for Windows x64/ARM64, macOS Apple Silicon/Intel, Linux x64/ARM64
- SHA-256 sidecars and schema-2 manifest for every published binary
- install.ps1 and install.sh verify downloads against sidecar and manifest
- cli/stable and cli/alpha release channels with an in-app updater
- Windows console multi-line paste folds into a single paste chip
- /model list refreshes when the model catalog arrives after connecting
- Interactive terminal UI with slash commands and keyboard shortcuts
- Headless one-shot and scripted runs with --disallowed-tools support
- Bounded turns capped at 24 base model calls and 600 seconds
- Custom agents, skills, plugins, hooks and MCP server support
- Plan Mode, subagents, background tasks and an agent dashboard
About RightNow AI
RightNow Agent is a command-line coding harness for developers who'd rather stay in the terminal than open an IDE. It's written in Rust and forked from xAI's grok-build, and the vendor's own measurements put session opens at 105-115 ms with a reused warm agent service against 631-728 ms without one. Dynamic tool sending dropped prompt tokens from 9,676 to 1,976 across three headless create-file runs, and one-call transactions finish a bounded create, edit or named-test straight from the typed tool result plus an objective postcondition instead of burning a second narration call. The launch story changed on 2026-09-09. Release 1.0.4 shipped native binaries for Windows x64, Windows ARM64, macOS Apple Silicon, macOS Intel, Linux x64 and Linux ARM64, each with SHA-256 sidecars and a schema-2 manifest, plus install.ps1, install.sh, an in-app updater and cli/stable and cli/alpha channels. Version 1.0.5 landed the same day with Windows console multi-line paste folding, a /model list that refreshes when the catalog arrives late, and publish-signed Windows ARM64 builds. The editor release line closed on 2026-09-05, leaving the agent as the sole product. You drive it through an interactive TUI or headless one-shot mode, with custom agents, skills, plugins, hooks and MCP servers supported and bounded turns capped at 24 base model calls and 600 seconds. GPU work carries over from the earlier line: CUDA, Triton, Mojo, PyTorch and Numba profiling, emulation and benchmarking across 86+ GPU architectures, with no physical GPU required. Where it sits: sub-$30 CLI agents like Claude Code and Codex bundle a frontier model, while RightNow Agent makes you bring a key to RunInfra or your own provider. The trade is control over the harness and signed, verifiable binaries over a bundled model subscription. It is YC-backed and SOC 2 Type 2 compliant.
Behind the Verdict
Pick RightNow Agent if you already live in a terminal and you want the harness, not the model subscription. The billing model is the tell: the coding plan starts at $10/mo (Starter, roughly $12 of usage, with 5-hour and weekly limits and a free standby tier), but your model tokens come from a separate prepaid RunInfra balance — $5 minimum top-up, credits valid a year, per-model prices like $0.05 input / $0.01 cached input / $0.15 output per 1M tokens on Nemotron 3.5 Lightning 30B. Budget both lines. Where it earns its place is speed and verifiability. Signed binaries with SHA-256 sidecars mean an updater that actually checks what it's installing. Headless one-shot runs with --disallowed-tools mean you can put it in scripts and CI without a human watching a chat window. The one-call transaction model cuts agent round-trips in half on bounded edits, which is the difference between a two-second and a six-second wait when you're doing twenty of them an hour. Pass if you want an all-in-one. Claude Code and Codex hand you a frontier model plus the harness for one monthly fee, and their context handling is tuned for long multi-file refactors. RightNow Agent is not trying to be that, and its measured wins are on short, bounded tasks — the token-saving numbers come from three headless create-file runs, not from a twelve-hour session. The GPU side deserves separate consideration. Forge CLI's swarm generates CUDA and Triton kernels with 32 Coder+Judge pairs, a MAP-Elites optimizer over 36 behavior cells and pattern RAG over 1,711 CUTLASS plus 113 Triton patterns. If you're hand-tuning kernels, that's a different job than autocomplete, and there's no general-purpose CLI agent competing on it. The 5x faster than torch.compile and 97.6% correctness claims are the vendor's —
Researching RightNow AI? 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 RightNow AI actually fits — and what changes day-one when you adopt it.
Installs via install.sh, which verifies the download against the SHA-256 sidecar and schema-2 manifest, then runs forge browse to pick a KernelBench task and generates a Triton variant, benchmarking it against torch.compile.
Outcome: A drop-in Triton or CUDA kernel with profiling evidence, plus a rollback directory and a recorded installer.toml if the version needs reverting.
Drives the agent headless on a create-file task and checks the bounded transaction result rather than parsing narrative output, using --disallowed-tools to keep the toolset narrow.
Outcome: Edits that either satisfy the objective postcondition or get rejected, with no second model call spent on narration.
Sets a Groq or Cerebras key and points the agent at a provider other than the shipped RunInfra default, tuning reasoning_effort where a model publishes a required value for JSON mode.
Outcome: The same harness behaviour on a provider the team already pays for, without re-learning the terminal workflow.
Use Cases
- Windows, macOS or Linux developer who wants a fast terminal coding agent with sub-300ms harness overhead per turn
- Developer who wants scriptable agent edits with objective postconditions instead of narrated diffs
- GPU kernel developer generating CUDA or Triton kernels through Forge CLI's swarm search
- HPC engineer profiling and emulating PyTorch kernels across 86+ GPU architectures without owning the hardware
- Coder who brings a Cerebras, Groq, OpenRouter or Vercel AI Gateway key and wants provider choice
- Team that wants a signed, manifest-verified binary with rollback (.old) and completion installs
Models Under the Hood
as of 2026-10-18
Limitations
- Docs state RightNow Agent, the terminal coding agent, is paused, with its guides kept online as reference.
- It runs approved tools on your machine and sends prompts and relevant tool results to the selected model provider via an API key.
- The 105-115 ms session-open figures are dated 2026-08-31 Windows observations from the repository, not performance guarantees, and the recorded external wall time was worse and remained unresolved.
- Turn execution is bounded, and paid Model APIs carry a $5 minimum top-up with credits lasting a year.
as of 2026-10-07
Verification history
We have re-verified RightNow AI 10 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 10 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.
Plans compared
For each published RightNow AI tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free Standby
$0/mo
Starter
$10/mo
Ideal for
One developer trying the harness on real work before committing — about $12 of usage a month with a $1.11 5-hour limit and $2.77 weekly limit.
What this tier adds
Starting paid tier: a 5-hour and weekly allowance plus Free Standby of up to $0.20 of chat a day when you hit a limit.
Pro
$29/mo
Ideal for
A working developer running daily agent sessions and kernel profiling, at roughly $35 of usage a month.
What this tier adds
Roughly triple Starter's allowance — $3.23 per 5 hours and $8.08 weekly, with Free Standby up to $0.58 a day and 5 to 32 concurrent requests per model.
Team
$99/mo
Ideal for
A small group sharing one workspace budget rather than paying per seat, at about $119 of usage a month.
What this tier adds
The largest self-serve allowance — $10.98 per 5 hours, $27.46 weekly, up to $1.98 a day of Free Standby and 8 to 32 concurrent requests per model.
Model APIs Pay as you go
From $5 top-up
Enterprise
Custom
Ideal for
Organizations that need private infrastructure, reserved capacity and a contractual SLA up to 99.99%.
What this tier adds
Adds audit logs and role-based access control, custom volume and contract terms, SOC 2 Type II attestation, a dedicated CSM and a private Slack channel.
Where the pricing makes sense
The company stage and team size where RightNow AI's pricing actually pencils out — and where peers do it cheaper.
The coding plans run $10/mo Starter (about $12 of usage), $29/mo Pro (about $35) and $99/mo Team (about $119), with self-serve credit top-ups from $5 for Model APIs. Team is shared per workspace, not per seat, so a small group is cheap relative to per-seat terminal agents. Below that, RunInfra Model APIs price per token — Nemotron 3.5 Lightning 30B at $0.05 input/$0.15 output per 1M tokens is the cheapest listed model.
Setup time & first value
How long it actually takes to get something useful out of RightNow AI — broken out by persona, not the marketing-page minute.
On Linux or macOS, install.sh verifies the sidecar and manifest and installs into ~/.rightnow/bin with completions and PATH updates — expect a few minutes end to end. On Windows, install.ps1 does the same job. From there, setting a provider key and running the quickstart is minutes; the TUI honours --disallowed-tools, and `rightnow doctor` is the documented first stop if something is off.
Switching to or from RightNow AI
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Claude Code or Codex: the docs include a dedicated 'Migrate from Claude Code or Codex' page plus configuration reference for project rules and permissions.
- →From the RightNow Editor line: the 2026-09-05 checkpoint ended the editor line, so prior Forge and editor workflows exist only as changelog history, not a supported path.
- ↗To Claude Code or Codex: config maps back through the same agent/skills/MCP concepts, but Forge's swarm kernel generation has no direct equivalent.
- ↗To a local-model setup: the shipped version 6 configuration removed the Ollama, llama.cpp, local weights and local embeddings paths, so this now needs a different tool.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “RightNow AI”, and we withheld 6: 6 could not be judged, because “RightNow AI” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about RightNow AI.
Official links
Tools that pair well with RightNow AI
Common stack mates teams adopt alongside RightNow AI, with the specific reason each pairing earns its keep.
Bito
Bito Governor is an AI model router with a code context engine that cuts coding-agent spend by 40-70%.
Claude Code
Anthropic's agentic coding assistant that plans, edits, and runs commands across your repo from the terminal, IDE, browser, or cloud session.
Windsurf
Devin Desktop (formerly Windsurf) is an agent-native IDE for running fleets of local and cloud coding agents from one Sessions board.
Featured Head-to-Head Comparisons
Rightnow Ai vs Spider Cloud
Choose RightNow AI if you're a GPU kernel developer needing a specialized IDE with GPU emulation, real-time NCU profiling, and AI autocomplete for CUDA/Triton. Choose Spider Cloud if you're building AI agents or RAG pipelines that require fast, reliable web scraping at scale with structured output. They solve completely different problems — one is for writing GPU kernels, the other for fetching web data.
Rightnow Ai vs Temporal Ai
If you're a GPU kernel developer needing a specialized IDE with emulation and profiling for CUDA/Triton, RightNow AI is the clear choice—especially given its recent agent-based kernel generation achieving major speedups. If you need reliable orchestration for AI agents or microservices with automatic retries and state persistence, Temporal AI is the mature, open-source platform used by major companies like OpenAI. They solve fundamentally different problems; choose based on whether your bottleneck is GPU optimization or workflow reliability.
Rightnow Ai vs Voyage Ai
Do not buy both unless you have unrelated needs. If your focus is high-accuracy RAG on specialized domains like finance or legal, choose Voyage AI for its tailored embedding models and rerankers. If you are optimizing CUDA/Triton kernels for NVIDIA GPUs, RightNow AI is the only dedicated AI-powered IDE with profiling, emulation, and multi-DSL support. For mixed workloads, consider hybrid workflows using both tools separately.
Pieces For Developers vs Rightnow Ai
If you write CUDA/Triton kernels and need AI-aided profiling, benchmarking, and code generation, RightNow AI is a no-brainer. If your pain is losing context across apps and needing an automatically searchable history of your work, Pieces for Developers is the pick. They solve fundamentally different problems, so choose based on whether you optimize GPU code or your personal workflow.
Bito vs Rightnow Ai
RightNow AI is your pick if you write GPU kernels and need integrated profiling, emulation, and AI optimization. Bito wins if your team relies on AI coding agents and struggles with cross-repo dependencies, architectural planning, or onboarding. If you do both, consider both—but for most, the choice reduces to: GPU performance or system-wide context?
Alternatives to RightNow AI
View all
Bito
Bito Governor is an AI model router with a code context engine that cuts coding-agent spend by 40-70%.
Claude Code
Anthropic's agentic coding assistant that plans, edits, and runs commands across your repo from the terminal, IDE, browser, or cloud session.
Frequently Asked Questions
Categories
Best-of guides
Used RightNow AI? Help shape our editorial sentiment research.