DeepWideResearch
Open-source agentic research tool where depth and width parameters control how far and how broad each query runs.
DeepWideResearch earns its place when research scope is a knob you need to turn. The deep/wide parameters, MCP source plugging, MIT-licensed code, and model flexibility give you control that closed deep research products do not expose — the vendor's own comparison page puts it against GenSpark, OpenAI, Manus, Gemini, Jina, and LangChain on exactly those axes. Credit metering on the cloud tiers (100 credits on Free, 2,000 on Plus at $15/month, 15,000 on Pro at $100/month) and the setup effort of running your own infrastructure are the real costs to weigh. If you want a one-click report and nothing else, use Gemini or OpenAI instead.
Verified 3d ago · liveness 67/100 · cite: rightaichoice.com/tools/deepwideresearch
- Research teams that need to tune scope per query rather than accept a fixed report length
- Developers wiring research into their own apps through the API or MCP servers
- Organizations that must keep research data on their own infrastructure
- Teams that want to swap between OpenAI, Claude, and open-source models
- Non-technical users who want a one-click research report with zero configuration
- Teams that need native mobile or desktop apps
- Groups without anyone to run and maintain self-hosted infrastructure
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 DeepWideResearch if nobody on your team can read a cURL or Python example, configure an MCP server, or maintain self-hosted infrastructure — you would be paying to set up a system you cannot operate.
Cloud research is metered in credits, so a habit of running high-depth, high-width queries will burn a 100-credit Free month and then a 2,000-credit Plus month faster than the calendar suggests.
The cloud tiers are priced for small technical teams: Free at $0/month with 100 credits to evaluate, Plus at $15/month with 2,000 credits and API access for a solo developer or two-person research pod, Pro at $100/month with 15,000 credits for a team running regular jobs. Enterprise is quoted per deal. Compared with credit-metered peers like Tavily or Exa you are buying the full research loop, not just retrieval; compared with OpenAI or Gemini deep research you are buying configurability
In short
DeepWideResearch — Open-source agentic research tool where depth and width parameters control how far and how broad each query runs. Best for Research teams that need to tune scope per query rather than accept a fixed report length, Developers wiring research into their own apps through the API or MCP servers, Organizations that must keep research data on their own infrastructure. Free to start; paid plans from $15/mo.
What's new in DeepWideResearch
Checked 3 days agoAcross the latest 3 updates: 2 feature updates and 1 pricing change.
Introduced cloud-hosted plans
DeepWideResearch launched Free, Plus, Pro, and Enterprise cloud plans with credit-based usage, complementing self-hosting under the MIT license.
Added Anthropic Claude support
DeepWideResearch now supports Anthropic Claude as an AI model option alongside OpenAI and open-source models.
Released MCP integration
Integrated Model Context Protocol to connect external data sources and tools such as Notion, Exa, and Tavily.
Viability Score
How well maintained and how widely used is DeepWideResearch? 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
- Depth parameter to control how thoroughly one topic is investigated
- Width parameter to control how many adjacent topics are explored
- Agentic RAG for quick Q&A through to multi-turn deep analysis
- Model Context Protocol (MCP) support for external data sources and tools
- Documented MCP sources: Notion, Exa, Tavily, Context 7
- Model flexibility across OpenAI, Anthropic Claude, and open-source models
- REST research endpoint callable with cURL, Node.js, Python, and bash
- Streaming research responses returned over the API
- Self-hosting available under the MIT license
- Local knowledge base support when self-hosted
- Cloud-hosted plans metered by credits
- Custom integrations on the Pro tier
- Enterprise custom deployment, SLA and compliance
- Enterprise priority email and Slack support with a dedicated solutions architect
- Community support via GitHub Issues and Discord
About DeepWideResearch
DeepWideResearch is an open-source agentic research platform from PuppyAgent where every investigation is tuned by two numbers: depth (how hard it digs into a single topic) and width (how many adjacent topics it fans out to). You set both, then run the same question from a quick Q&A up to a full analysis without swapping tools. It connects external sources and tools through the Model Context Protocol — Notion, Exa, Tavily, and Context 7 are documented — and accepts multiple model backends including OpenAI, Anthropic Claude, and open-source options. Everything is callable from an API or SDK, with cURL, Node.js, Python, and bash examples published on the site; streaming responses come back over the API. The codebase runs under the MIT license, so you can self-host the whole platform with local knowledge bases and keep private data on your own infrastructure. If you would rather not run servers, PuppyAgent sells cloud-hosted plans metered by credits: Free at $0/month with 100 credits, Plus at $15/month with 2,000 credits plus API and SDK access, Pro at $100/month with 15,000 credits, and an Enterprise tier priced on request with unlimited credits, SLA, and 24/7 support. It is built for developers and research teams who want to wire research into their own stack rather than rent a closed black box.
Behind the Verdict
The interesting thing about DeepWideResearch is that it treats research scope as an input rather than a product SKU. Most deep research tools ship one behaviour — you hand over a question and get whatever length and breadth the vendor decided on. Here the API call carries a deepwide object: {"deep": 1.0, "wide": 1.0}, and moving those numbers up or down moves you along the range from quick Q&A to comprehensive analysis. For a competitive-intel analyst who wants twenty adjacent markets touched lightly this week and one market torn apart next week, that is a real workflow difference. Second, the plumbing is open. The platform speaks the Model Context Protocol and the vendor documents Exa and Tavily for web search, Context 7 for dev search, and Notion as a source. That means the retrieval layer is something you assemble from parts you choose, not a fixed index. It also means the quality of your output tracks the quality of the sources you wire in — a real dependency, and worth saying plainly. Third, ownership. The code is MIT-licensed and self-hostable, so private data can stay on your infrastructure with a local knowledge base and no third party in the loop. Against OpenAI or Gemini deep research, that is the sharpest differentiator the tool has. Against LangChain, the pitch is that the research loop is already built rather than a toolkit you assemble. Where it costs you: setup. This is a developer-shaped product. You will be reading API examples in cURL, Node.js, Python, or bash, configuring MCP servers, and making model choices across OpenAI, Anthropic Claude, and open-source backends. There is no native mobile or desktop app, and the Free cloud tier's 100 credits will not survive many deep runs. The Plus tier at $15/month gives 2,000 credits and API access; Pro at $100/month gives 15,000 credits and custom integrations. Enterprises get unlimited credits, SLA, custom deployment, priority Slack and email support, and a dedicated solutions architect — priced on request. Bottom line: excellent fit for a technically staffed research or platform team that wants research as a component. Poor fit for someone who wants a report, not a system.
Researching DeepWideResearch? 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 DeepWideResearch actually fits — and what changes day-one when you adopt it.
Calls the research endpoint from a Node.js backend with deep set low and wide set low for a fast answer on the user's question, then raises both values for a 'deep dive' button in the UI, streaming the response back to the browser.
Outcome: Ships a two-mode research feature without building a retrieval pipeline, using the Plus tier's 2,000 credits/month and API access.
Configures Exa and Tavily as MCP search servers plus a Notion workspace holding prior research, then runs a wide query across six adjacent competitors at depth 0.5 before drilling into the two that matter at depth 1.0.
Outcome: One pass produces a landscape map and a deep dossier on the finalists, from a single tool with consistent sourcing.
Self-hosts DeepWideResearch under the MIT license on internal infrastructure, points it at a local knowledge base and an internal docs MCP server, and selects an open-source model so no query text leaves the network.
Outcome: Internal research workflows stay inside the perimeter while still using the depth/width agent loop.
Use Cases
- Run the same market research question at low depth/width for a fast answer, then re-run it wider for a full landscape scan.
- Build an agentic RAG pipeline against a private knowledge base for internal discovery.
- Aggregate results from Exa and Tavily in one research pass instead of querying each separately.
- Connect Notion through an MCP server so research pulls from your own team's pages.
- Self-host a private research assistant for sensitive internal data under the MIT license.
- Benchmark different model backends (OpenAI, Claude, open-source) against the same query.
Models Under the Hood
as of 2026-10-10
Limitations
- Cloud plans are credit-metered — 100 credits/month on Free, 2,000 on Plus at $15/month, 15,000 on Pro at $100/month, unlimited on Enterprise — so heavy research schedules need sizing before you commit.
- The research endpoint takes a deepwide object with deep and wide values, which means integrating it requires reading API examples and configuring MCP servers rather than clicking a button.
- Self-hosting is fully open source under MIT (v0.1.0) but you run and maintain the infrastructure.
- Output quality tracks the quality of the MCP sources you connect, such as Exa, Tavily, Context 7, and Notion.
as of 2026-10-08
Verification history
We have re-verified DeepWideResearch 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-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
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.
Plans compared
For each published DeepWideResearch tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0/mo
Ideal for
A developer or evaluator who wants to test the research endpoint and the depth/width controls before spending anything, at $0/month.
What this tier adds
Starting tier: 100 credits/month and community support, but no API access.
Plus
$15/mo
Ideal for
A solo developer or small team putting research into a real product or recurring workflow, at $15/month.
What this tier adds
Adds API and SDK access plus 2,000 credits/month versus Free, opening programmatic use.
Pro
$100/mo
Ideal for
A team running regular, higher-volume research jobs that outgrow 2,000 credits, at $100/month.
What this tier adds
Raises the allowance from 2,000 to 15,000 credits/month and adds dedicated support and custom integrations.
Enterprise
Custom
Ideal for
Organizations needing unlimited usage, compliance commitments, or deployment on their own terms; priced on request.
What this tier adds
Custom pricing with unlimited credits, 24/7 support, SLA and compliance, custom deployment, and a dedicated solutions architect.
Where the pricing makes sense
The company stage and team size where DeepWideResearch's pricing actually pencils out — and where peers do it cheaper.
The cloud tiers are priced for small technical teams: Free at $0/month with 100 credits to evaluate, Plus at $15/month with 2,000 credits and API access for a solo developer or two-person research pod, Pro at $100/month with 15,000 credits for a team running regular jobs. Enterprise is quoted per deal. Compared with credit-metered peers like Tavily or Exa you are buying the full research loop, not just retrieval; compared with OpenAI or Gemini deep research you are buying configurability
Setup time & first value
How long it actually takes to get something useful out of DeepWideResearch — broken out by persona, not the marketing-page minute.
Cloud: about 15–30 minutes for a developer to get an API key and a first cURL or Python call returning a streamed research result, longer if MCP servers need configuring. Self-hosted: allow a half day or more — cloning the MIT-licensed repo, standing up the service and its dependencies, wiring MCP servers for Exa, Tavily, or Notion, and connecting a local knowledge base. Non-technical users
Switching to or from DeepWideResearch
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From a manual search-and-notes workflow: replace ad-hoc browser research with an API call that sets depth and width per question.
- →From a closed deep research tool: reproduce the report you get today, then raise width to add adjacent topics the closed tool never covered.
- →From a DIY LangChain retrieval pipeline: keep your sources but drop the custom agent loop in favour of the built-in deep/wide parameters and MCP connectors.
- →From parallel Exa and Tavily scripts: connect both as MCP servers and let one research request query across them.
- ↗To a closed deep research product: you lose depth/width control, self-hosting, and model swapping, and gain a one-click report with no infrastructure.
- ↗To a retrieval-only API: you keep search control but rebuild the reasoning and synthesis loop yourself.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “DeepWideResearch”, and we withheld 6: 6 could not be judged, because “DeepWideResearch” 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 DeepWideResearch.
Official links
Tools that pair well with DeepWideResearch
Common stack mates teams adopt alongside DeepWideResearch, with the specific reason each pairing earns its keep.
Anara
Anara is an AI research workspace that searches your library and the web and cites every answer back to the exact passage in the source.
Imbue
Imbue is an open AI lab publishing modular, open-source coding-agent tools you run and inspect yourself.
Buildpad
Buildpad runs parallel AI market research agents to validate your startup idea before you build.
Featured Head-to-Head Comparisons
Deepwideresearch vs Truleo
If you run a law enforcement agency and need to unify siloed data (RMS, jail calls, body cameras) with automated leads and report writing, Truleo is purpose-built for you. For technical researchers or developers who want granular control over research scope, multi-engine search, and self-hosting, DeepWideResearch offers unmatched flexibility at a lower starting cost.
Deepwideresearch vs Praktika
If your goal is to improve spoken fluency through AI conversation practice, Praktika is the right choice. If you need an open-source, configurable research tool to analyze large amounts of data across multiple sources, DeepWideResearch is better. They serve entirely different needs.
Deepwideresearch vs Presto Voice
Presto Voice and DeepWideResearch serve entirely different domains. Presto Voice is a specialized drive-thru AI for QSR chains aiming to boost revenue through automation and upselling, with a proven 95% non-intervention rate. DeepWideResearch is a flexible, open-source research agent for technical users who need granular control over depth and width of analysis. Choose Presto Voice if you run a QSR chain; pick DeepWideResearch for customizable, self-hosted research.
Alternatives to DeepWideResearch
View all
Frequently Asked Questions
Used DeepWideResearch? Help shape our editorial sentiment research.