The guidebook

Everything SIH 2026 asks of you

Written for a team entering for the first time. Every rule here is quoted from an official source with the link beside it, or counted from the 229 published problem statements when you loaded this page. You have 9 days until ideas close on 15 Sept.

Read the 229 problem statements

What SIH actually is

The shape of the thing, before any of the rules make sense.

Smart India Hackathon is run by the Ministry of Education. Government ministries, departments and public bodies each publish problems they actually want solved — this year, 229 of them — and student teams pick one and propose a solution.

It happens in three parts, and almost every first-year only knows about the third. First your college runs its own internal hackathon and forwards a limited number of teams. Then those teams file an idea document against a specific problem statement on sih.gov.in. Only then, months later, do the shortlisted teams travel to a national grand finale and build for 36 hours.

The part that decides your year is the writing, not the building. You are eliminated or shortlisted entirely on a document, before a line of code exists.

The road from here to the finale

What happened to everybody who entered in 2025. The last two lines are counted from the published result tables; the first three are reported participation totals, which is press coverage of a government event rather than the government.

Teams in college internal hackathons
1,42,715
Reported participation figure for SIH 2025 · reported
Teams their college nominated
68,766
Roughly half the field is out before the national round begins · reported
Ideas submitted
72,165
Across 271 problem statements — about 266 each, against a cap of 500 · reported
Shortlist notices
1,870
Counted from the four published screening batches
Prize-winning entries
321
Counted from the grand finale result table

Half the field is out before the national round begins, and only about one idea in 39 was shortlisted. Both gates close before you write any code, which is why this whole site is about choosing rather than building.

Shortlist and winner counts parsed from the 2025 screening results, read 27 August.

Whether your team can enter at all

Quoted from the portal's own FAQ. Every one of these is a thing you cannot fix in September — you cannot add a sixth member after your college has nominated, and you cannot enter at all if your college never ran its internal hackathon.

Six members, including the team leader
"Each team must consist of 6 members, including the team leader"
At least one member must be female
"must include at least one female member"
Your college has to run an internal hackathon first
"it is mandatory for all the participating colleges/institutes to organise intra college hackathons"
A college nominates 45+5 teams; a university, 100
"HEI may be allowed to nominate 45+05 teams/colleges … in case of university 100 Team"

Quoted from the Official SIH FAQ, read 29 August.

What you are choosing between

All 229 of them, counted when you loaded this page. This used to be a separate page called "By the numbers"; it belongs here, between whether you can enter and how you are scored.

How many a software team with no budget can really enter

124 of 229

The other 105 need something sourced before you can start — components, or data nobody has given you. Not worse problem statements, a different commitment: our reading of each brief, with hardware and the missing dataset taken from the organisers' own fields.

Software or hardware, in the organisers’ own words

175

Software76%Hardware24%

34: open Student Innovation themes
44: arrive with a dataset
5: come with a briefing video

What you would actually be building

Our reading of the parts each brief asks for, not the two-bucket column. This is what decides whether your team has the right people.

ML-heavy
96 · 42%
Hardware-integrated
54 · 24%
Dashboard
37 · 16%
Not clear from the brief
27 · 12%
Mobile / offline
15 · 7%

Which areas are crowded before anybody has entered

The organisers’ own 17 theme labels, largest first. The sizes are wildly uneven, and your odds are slots divided by how many teams pick the same statement.

Smart Automation
55 · 24%
Blockchain & Cybersecurity
31 · 14%
Disaster Management
25 · 11%
Agriculture, FoodTech & Rural Development
22 · 10%
MedTech / BioTech / HealthTech
18 · 8%
Miscellaneous
15 · 7%
Robotics and Drones
12 · 5%
Smart Education
11 · 5%
Space Technology
9 · 4%

Who sets these problems, and therefore who marks them

32 organisations. The largest sets 34 of the 229 on its own. 11 of them have no prior record in the published 2024 or 2025 result tables — which is a gap in what we know, not a finding about them, and nothing on this site infers what an evaluator wants from which organisation they are.

AICTE
34 · 15%
Ministry of Earth Sciences (MoES)
30 · 13%
National Technical Research Organisation (NTRO)
23 · 10%
Ministry of Home Affairs
11 · 5%
Indian Space Research Organisation(ISRO)
11 · 5%
Ministry of Rural Development
10 · 4%
Ministry of Consumer Affairs, Food & Public Distribution
10 · 4%
Government Of Maharashtra
9 · 4%

How much the briefs leave unsaid

77 of the 229 have no “Expected Solution” section at all, and 34 never state a deliverable because they are not meant to. A brief with no stated deliverable is not unenterable — it means the jury has more discretion, so what you choose to build carries more of the weight than usual.

Your real odds, and why the problem statement decides them

Division, not a prediction. Move the slider to the number of ideas you expect on the one you want.

this is the whole argument for choosing carefully

Ideas filed against the problem statement you want

266 of 500

Your chance of a shortlist notice

2.6% · about 1 in 38

Getting loud. You need a differentiator, not just a solution.

7 shortlist slots divided by the number of ideas filed against that problem statement. The portal publishes that number against every one of them, so you can redo this yourself at any point.

The guidelines say four to five teams per problem statement reach the finale. The four published 2025 screening batches actually ran three to eight, with 7 by far the most common, so that is the number used here and everywhere else on this site.

Counted from the 2025 screening results, read 27 August.

What you submit, and in what shape

The section headings of the organisers' own idea template, in submission order. Unchanged across 2024, 2025 and 2026.

01Title / Idea title
02Proposed Solution
03Technical Approach
04Feasibility and Viability
05Impact and Benefits
06Research and References

It is filed as a PDF built from the organisers’ template, not as a slide deck of your own design.
The portal asks for three things, not one: idea title, idea description, idea presentation (pdf). Everybody prepares the presentation; almost nobody writes the title and the description before the night they file.
You submit against 2 problem statements, not one. Two entries, two decisions — and each one freezes at 500 submissions.
Your solution must not have appeared in any previous event or programme, of any sort. That is the clause that catches a recycled college project, which is exactly what a team under time pressure reaches for.
The organisation that set a problem statement is not obliged to declare a winner for it. It is rare, and we do not turn it into a warning on any individual problem statement.

Four things you cannot undo

Team details cannot be changed once the SPOC has entered them. "Team details once entered cannot be altered"
Your team name must not contain your institute's name in any form. "The team name should be unique and must not contain the name of your institute in any form"
The authorisation letter needs your principal's signature and the college seal, on letterhead. "The letter must be issued on college letterhead only... duly signed by college principal/ dean/ institute director and bear the college seal. This letter will be required during the grand finale"
Your team leader only gets portal credentials after your SPOC uploads the team. "Nominated Team's team leader will get the Login credentials after SPOC upload team details in the portal to complete the idea submission process"

From the official idea presentation template, read 27 August.

The organisers publish two last dates

The problem-statement table gives 30 September 2026 against every one of the 229. The College SPOC letter gives 15 September 2026, and adds that no request will be entertained after it. The same page of that letter also gives 30 August for the SPOC’s own step, which has passed — so neither document can simply be declared the winner.

We have not resolved it and we are not going to guess. Every countdown on this site runs to 15 September 2026, because the two are not equally risky to be wrong about: a team that plans to the earlier date and finds the portal still open has five days of slack, and a team that plans to the later one and meets a cutoff on the 15th has filed nothing.

the "Deadline for Idea Submission" column against every problem statement — 30 September 2026.
"The last date for team nomination and idea submission by College SPOC and Team leader on SIH portal is till 15th Sept 2026 only. No request will be entertained after the deadline" — 15 September 2026.

The irreversibles and the second last date, from the 2026 College SPOC guidelines, read 31 August.

The nine things you are scored on

The evaluation parameters in the official guidelines. The same nine for every problem statement, which makes them the closest thing to a rubric this competition publishes.

01Novelty of the idea
02Complexity of the solution
03Clarity and detail, in the prescribed format
04Feasibility
05Practicability
06Sustainability
07Scale of impact
08User experience
09Potential for future work

Two of these are about the writing rather than the idea. “Clarity and detail, in the prescribed format” means a document that invents its own headings is losing marks on a stated parameter before anybody reads what you built. And “potential for future work” rewards saying what you would do next, which most teams leave out entirely.

From the SIH 2026 Guidelines (PDF), read 27 August.

Questions people actually ask

The ones that get answered wrong by seniors. Open on paper — everything here prints.

I am in my first year. Can I even enter?

SIH is open to students at participating colleges, and first-years do enter and do reach the finale. What decides it is your college's internal hackathon, not your year — the eligibility rule by year of study is one we have not been able to quote from an official source, so ask your SPOC rather than taking anyone's word for it, including ours.

Our team leader has no login for sih.gov.in. Is something wrong?

Possibly, and it is worth chasing today. The letter says the team leader gets login credentials only after the SPOC uploads the team details on the portal — so no login means your team has probably not been uploaded yet. There is no notification either way, and the SPOC's own nomination window closes before the team leader's submission window, so a team that notices this on the last day has already lost. Ask your SPOC to confirm your team is uploaded. The other two explanations are that your team was not among the 50 your college may nominate (45 shortlisted and 5 waitlisted), or that your college never registered a SPOC or ran an internal hackathon at all. Something like this is enough to send: “Respected Sir/Ma’am, our team was selected in our college’s internal hackathon for SIH 2026. Our team leader has not yet received login credentials for sih.gov.in. As the credentials are issued only after the SPOC uploads the team details on the portal, could you please confirm whether our team has been uploaded? The idea submission has to be completed by the team leader from that login.”

What is a SPOC, and why does everyone keep saying it?

The Single Point of Contact: the faculty member your college has registered as its SIH coordinator. They run the internal hackathon, they submit your college's nominations, and they are the person who knows the dates. Finding out who yours is should be the first thing your team does.

Does my college have to run an internal hackathon?

Yes. The FAQ is explicit that it is mandatory for participating institutes to organise intra-college hackathons, and a college nominates its teams out of that. If yours has not announced one, that is the thing to chase — you cannot route around it.

Do we have to build the whole thing before we submit?

No. The idea submission is a document. You are shortlisted or eliminated on it, and the 36 hours of building only happen at the grand finale, months later, if you get there.

Can we submit the project we already built for a college fest?

The guidelines say the solution must not have appeared in any previous event or programme of any sort. A recycled project is exactly what a team under time pressure reaches for, and it is exactly what that clause exists to catch.

How do we pick between two problem statements?

You do not have to. You file against two, and they should not be the same bet twice — two hardware statements from the same ministry means the same components to source and the same evaluators reading both. One stated problem and one open theme is genuinely two chances. And 7 shortlist slots divided by the number of teams who picked the same statement is the maths that actually moves your odds.

What is an 'open' problem statement?

34 of the 229 this year are open Student Innovation themes. They name an area and nothing else: there is no requirement list to satisfy, so your idea is judged entirely on its own merits. That cuts both ways — more freedom, and nothing to point at when a jury asks why you built what you built.

What if the problem statement needs data nobody gave us?

Only 44 of the 229 come with a dataset attached. For the rest, finding or generating data is part of the work and nobody tells you that up front. It is sometimes a weekend and sometimes impossible, and it is much cheaper to find out now than in week three.

Does it matter which ministry set the problem?

Not in the way people assume. The name of a ministry tells you nothing about what its evaluators will check; the brief's own wording does. If it insists on offline use, a language, an accuracy target, or integration with an existing system, that is what gets tested. Nothing on this site infers a preference from an organisation's name.

Can a problem statement close before the deadline?

Yes. Each one freezes at 500 submissions, so a crowded one can stop being an option while you are still deciding. If your first choice is filling, file against it early rather than perfecting the document first.

How many people do we need, and does it have to be six?

Six, including the team leader, and at least one member must be female. Both are quoted from the FAQ above. Neither is fixable late: your college nominates teams, and a team that is short or non-compliant at that point simply is not forwarded.

What happens after we submit?

Your idea goes to the organisation that set the problem statement, and they screen the entries against it. Shortlisted teams are announced in batches and travel to a national grand finale to build. We do not have the 2026 dates for either step, so we are not going to invent them.

What we still will not tell you

Shorter than it used to be, and worth keeping honest rather than quietly filling in.

Eligibility by year of study. The full calendar — when your college’s internal hackathon has to happen, when screening results are announced, when the grand finale is. Prize amounts. What the finale itself involves hour by hour.

We do not have those in a form we can quote, so we are not going to paraphrase them from memory. Read them on the portal, from the organisers, in their own words.

Read the official rules on sih.gov.in

How we would spend the days you have left

This section is our opinion, not a rule. Nothing here comes from the organisers.

01
Find your SPOC this week
Your college's internal hackathon date decides everything else on your plan, and it is the one thing you cannot look up. Ask, get a date, and work backwards from it.
02
Pick the pair, not the favourite
You get two entries. Two hardware problem statements from the same ministry is one bet made twice. One stated problem and one open theme is genuinely two chances.
03
Write against the nine parameters, not against your idea
Most teams describe what they built. The document is scored on novelty, feasibility, scale of impact, user experience and potential for future work — so say something deliberate about each, under the organisers' own six headings.
04
Decide the questions the brief left open, on purpose
Most of these never say which languages, what accuracy counts as success, or who exactly uses the thing. A jury can still ask. Deciding deliberately and writing it down beats deciding by accident in the last week.
05
Check whether the data exists before you commit
A problem statement that needs historical records with no dataset attached is asking you to find them. That is sometimes a weekend and sometimes impossible.
06
Watch the counter on the ones you want
With about 7 slots per problem statement, every extra hundred ideas filed against your choice roughly halves your odds — and each one freezes at 500.
07
Build the smallest thing that demonstrates the claim
A working narrow demo of the thing the brief actually asked for reads better than a broad prototype that gestures at everything and finishes nothing.